U.S. patent application number 10/575443 was filed with the patent office on 2007-08-16 for fail-safe pneumatically actuated valve with fast time response and adjustable conductance.
This patent application is currently assigned to SUNDEW TECHNOLOGIES, LLC. Invention is credited to Ofer Sneh.
Application Number | 20070187634 10/575443 |
Document ID | / |
Family ID | 34465331 |
Filed Date | 2007-08-16 |
United States Patent
Application |
20070187634 |
Kind Code |
A1 |
Sneh; Ofer |
August 16, 2007 |
Fail-safe pneumatically actuated valve with fast time response and
adjustable conductance
Abstract
Apparatus and method for fail-safe high-speed-pneumatic valve is
disclosed. Fail-safe dependability is provided by a spring-loaded
normally-closed pneumatic actuator. When the spring-loaded actuator
is pressurized, the normally closed mechanism is actuated to the
valve active position. Concurrently, the pressure is directly
applied to deflect a diaphragm or a bellow-assembly back to sealing
position. Ultra high purity embodiments with standard dome shaped
diaphragms are disclosed. Additional high conductance diaphragms
and bellows embodiments are employed for higher conductance valves.
Novel flow path layouts are disclosed. The valves are applicable
for fast gas and fluid switching and are particularly suitable for
high productivity Atomic Layer Deposition (ALD) applications.
Additional embodiments cover improved diaphragm and seal
reliability, externally adjustable valve conductance, improved
valve safety and high temperature valve seals.
Inventors: |
Sneh; Ofer; (Boulder,
CO) |
Correspondence
Address: |
PATTON BOGGS LLP
1801 CALFORNIA STREET
SUITE 4900
DENVER
CO
80202
US
|
Assignee: |
SUNDEW TECHNOLOGIES, LLC
3400 Industrial Lane Unit 7
Broomfield
CO
80020
|
Family ID: |
34465331 |
Appl. No.: |
10/575443 |
Filed: |
October 18, 2004 |
PCT Filed: |
October 18, 2004 |
PCT NO: |
PCT/US04/34453 |
371 Date: |
December 18, 2006 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
60512236 |
Oct 17, 2003 |
|
|
|
Current U.S.
Class: |
251/30.01 ;
251/26 |
Current CPC
Class: |
F16K 31/003 20130101;
Y10T 137/87209 20150401; F16K 7/14 20130101; F16K 31/1221 20130101;
F16K 7/17 20130101 |
Class at
Publication: |
251/030.01 ;
251/026 |
International
Class: |
F16K 31/12 20060101
F16K031/12 |
Claims
1. A fluid control valve comprising: a valve seat; a flow path
through said valve seat; a diaphragm; a normally closed pneumatic
actuator; a valve control chamber; a pneumatic feed line; and a
pilot valve; wherein said diaphragm is dispersed between said valve
seat and said valve control chamber; said normally closed pneumatic
actuator is configured to normally close said flow path by
deflecting said diaphragm to seal over said valve seat; said
pneumatic feed line is in fluidic communication with said normally
closed pneumatic actuator; and said pneumatic feed line is in
fluidic communication with said valve control chamber through said
pilot valve.
2. The fluid control valve as in claim 1 wherein: said pilot valve
is a three way normally open valve; said control chamber
communicates with said pneumatic feed line through said pilot valve
when said pilot valve is not actuated; said control chamber is
disconnected from said pneumatic feed line by said pilot valve when
said pilot valve is actuated; and said control chamber communicates
with a vent line through said pilot valve when said pilot valve is
actuated.
3-9. (canceled)
10. The fluid control valve as in claim 1 wherein: said pneumatic
actuator includes a stem; said stem penetrates through the wall of
said valve control chamber; and a sliding seal is dispersed between
said stem and said wall of said valve control chamber.
11-78. (canceled)
79. A fluid control valve comprising: a valve body having an inlet
port and an outlet port; a valve chamber bottom portion formed in
said valve body wherein a first of said ports is connected in
serial fluidic communication into said valve bottom portion
substantially at the center of said valve chamber bottom portion; a
second of said ports is connected in serial fluidic communication
into said valve bottom portion substantially off the center of said
valve chamber bottom portion; a valve seal located inside said
valve chamber bottom around said first port; a valve chamber top
portion made from a substantially flexible member; the center of
said substantially flexible member is normally positioned
substantially separated from said valve chamber bottom portion;
said valve chamber top separates said valve chamber from a valve
control chamber; said valve control chamber comprises a fluid
connection port; said valve control chamber comprises a
translatable stem actuated by pressurized fluid through a fluid
feed line; a pilot control valve connected in serial fluidic
communication between said fluid feed line and said fluid
connection port; the fluid path in said pilot control valve
normally connects said fluid feed line to said fluid connection
port; said fluid path in said pilot valve disconnects said fluid
feed line from said fluid connection port when actuated and
connects said fluid connection port to a vent line when actuated;
said valve stem is normally compressed with a spring to push and
deflect said flexible member between said valve chamber and said
control chamber to conform and substantially seal over said valve
seal; wherein a fluid is applied through said fluid feed line to
actuate said valve stem to translate it away from said flexible
member between said valve chamber and said control chamber; said
fluid is applied into said valve control chamber through said fluid
path in said pilot valve when said pilot valve is not actuated to
deflect said flexible member between said valve chamber and said
control chamber to substantially seal over said valve seal; said
fluid is vented out of said valve control chamber through said vent
line when said pilot valve is actuated to permit said flexible
member to return to an undeflected position thereby opening said
fluid control valve.
80-138. (canceled)
139. A method of operating a fluid control valve comprising:
mechanically holding said valve closed in an inactive state in
which it cannot be operated pneumatically; changing said valve to
an active state in which it can be opened and closed pneumatically;
and opening and closing said valve pneumatically.
140. A method as in claim 139 wherein said changing comprises
pneumatically actuating a mechanical valve actuator.
141. A method as in claim 139 wherein said mechanically holding
comprises holding said valve closed with a spring.
142. A method of operating a fluid control valve comprising:
holding said valve diaphragm closed with a mechanical actuator;
releasing said mechanical actuator; and opening and closing said
valve diaphragm pneumatically.
143. A method as in claim 142 wherein said releasing is performed
pneumatically.
144. A method as in claim 142 wherein coordinated with said
releasing, pneumatic pressure is substituted for mechanical
pressure to hold said valve closed.
145-148. (canceled)
149. A fluid control valve comprising: a valve seat; a flow path
through said valve seat a flexible member; a pneumatic actuator; a
flexible member chamber; a flexible member chamber evacuation port;
and an evacuation line; wherein said flexible member is dispersed
between said valve seat and said flexible member chamber; said
pneumatic actuator is configured to close said flow path by
deflecting said flexible member to seal over said valve seat; said
flexible member chamber is pressure sealed; and said flow path
remains pressure sealed from the ambient when a flexible member
failure occurs.
150. The fluid control valve of claim 149, wherein said flexible
member comprising a metallic diaphragm.
151. The fluid control valve of claim 149, wherein said flexible
member comprising a metallic bellow.
152. The fluid control valve of claim 149, wherein said flexible
member chamber is further evacuated following a failure of said
flexible member.
153. A method of operating a fluid control valve, said method
comprising: providing a valve including a valve control chamber, a
valve seat, a fluid flow path through said valve seat, a valve
diaphragm, and a valve actuator; holding said valve diaphragm
closed with the force of said valve actuator in an inactive state;
and pneumatically reducing the force of said valve actuator against
said valve diaphragm while changing the pressure in said valve
chamber to hold said valve diaphragm closed to create an active
shut valve state.
154. A method as in claim 153 wherein said valve diaphragm is
located between said valve control chamber and said valve seat and
said changing the pressure in said valve chamber comprises
increasing the pressure in said valve chamber.
155. A method as in claim 153 wherein said providing further
comprises providing a piston connected to said valve actuator and
said pneumatically reducing comprises pneumatically forcing said
piston connected away from said valve diaphram.
156. A method as in claim 155 and further comprising releasing said
force on said piston to disable flow through said valve seat when
said diaphragm fails.
157. A method as in claim 153 and further comprising: releasing the
pressure in said valve control chamber to open said flow path
through said valve seat to create and active open valve state.
158. A method as in claim 157 wherein said providing further
comprises providing a pilot valve and said releasing comprises
venting said valve control chamber through said pilot valve.
159. A method as in claim 158 wherein said pilot valve is a
three-way normally open valve, said providing further comprises
providing a source of pressurized fluid, said changing the pressure
in said valve chamber comprises connecting said valve chamber to
said pressure source, and said releasing further comprises
actuating said pilot valve to disconnect said valve chamber from
said pressure source.
160. A method as in claim 159 and further comprising de-actuating
said pilot valve to connect said valve chamber to said source of
pressurized fluid to disable said fluid flow through said valve
seat.
161. A method as in claim 160 wherein the response time for said
disabling said fluid flow through said valve seat is one
millisecond or less.
162. A method as in claim 160 wherein the response time for said
disabling said fluid flow through said valve seat is one-half
millisecond or less.
163. A method as in claim 153 and further comprising adjusting the
conductance of fluid flow path through said valve seat, wherein
said adjusting is performed externally of said fluid valve.
164. A method as in claim 163 wherein said providing further
comprises providing a restricted gap between said valve actuator
and said valve diaphragm when said valve actuator is released, and
said adjusting comprises adjusting the travel of said valve
actuator thereby controlling the size of said restricted gap.
165. A method as in claim 153 and further comprising controlling
the pulsed delivery of gas into an atomic layer deposition (ALD)
apparatus using said fluid control valve.
Description
BACKGROUND OF THE INVENTION
[0001] 1. Field of the Invention
[0002] This invention relates to the field of fluid delivery and
more specifically to apparatus and method for switching reactive
and inert fluid with high speed and performance.
[0003] 2. Description of Prior Art
[0004] In the fabrication of semiconductor and similar devices,
substrates are being processed under conditions of controlled
ambient, which is accomplished within enclosed spaces, or chambers,
wherein fluids are delivered and exhausted. Shutoff valves are
commonly used to route the flow of fluids such as gasses and
liquids within fluid delivery manifolds. In particular,
diaphragm-based high purity and ultrahigh purity valves are
commonly used to command the flow of inert and reactive gas within
high purity manifolds that are kept under tight standards of
low-contamination levels.
[0005] Typically, a dome-shaped preformed metallic diaphragm is
implemented to create an all-metallic valve chamber over a valve
seat. The valve seat typically implements a polymeric seal and is
commonly located at the center of the valve chamber, across from
the diaphragm. The diaphragm is clamped at the perimeter and is
normally held at the unstressed dome shape. When the diaphragm is
not stressed, a fluid path is linked, through the valve seat, with
at least another fluid path that communicatively runs into the
valve chamber. Accordingly, the shutoff valve is "OPEN". To shut
the valve off, the diaphragm is deflected towards the valve seat by
a mechanical plunger to enclose the fluid path that runs through
the valve seat. A leak-free seal is accomplished with appropriately
selected valve seat material and matching sealing pressure applied
by the plunger over the diaphragm. When the stress is removed from
the diaphragm, the diaphragm flexes back to the dome shape to clear
the fluid path within the valve. The art of high purity diaphragm
valves includes a selection of different valve seat and diaphragm
designs that are proven useful for high-purity switching of fluids.
For example, U.S. Pat. No. 5,131,627 articulates several useful
methods to accomplish a reliable valve at high standards of
purity.
[0006] In the art of fluid control, the need exists to construct
fail-safe valves that are normally closed when the valve is not
energized. In particular, fail-safe valves are mandatory within
manifolds that are built to deliver hazardous or otherwise reactive
chemicals and gasses. Accordingly, normally closed diaphragm valves
are adapted with a spring-loaded plunger called the valve-stem. For
example, the disclosure in U.S. Pat. No. 5,131,627 accommodates the
valve stem and an energized spring within the valve bonnet.
Fail-safe diaphragm valves are actuated to open the fluid path when
the spring-loaded valve stem is pulled away from the diaphragm.
When these fail-safe valves are not actuated, they return to their
"normally-closed" position.
[0007] Automatic actuation of fail-safe normally-closed (FSNC)
valves is accomplished with a machine commanded actuator. The art
of machine commanded actuators includes pneumatic,
electromechanical, piezoelectric and electro-thermal
stem-actuation. Pneumatic-actuation has been the most widely
accepted method for machine commanded valve actuation due to its
superior reliability, safety and low cost. For example, a piston
type pneumatic actuator of a specific useful design is provided in
U.S. Pat. No. 5,131,627. Likewise, many other embodiments are
suggested within the prior art commonly with one or multiple
pistons that are arranged with a sliding seal within a matching
cylinder and are actuated when compressed fluid, typically air, is
communicated into the cylinder. The pressurized fluid applies force
on the sliding pistons to propel the motion of the pistons within
the cylinder. Typically, the valve stem is rigidly attached to the
pistons. The fluid is introduced to propel the pistons and the
attached valve stem to move away from the diaphragm. Commonly,
metallic diaphragms are pre-formed to a deformation-free state
wherein a gap exists between the mounted diaphragm and the valve
seat corresponding to an open valve. When the valve stem is removed
from the diaphragm, the diaphragm flexes, by its own elasticity,
back to the stress-free form. When the fluid is released from the
pneumatic actuator, the valve stem is returned by the force of the
energized spring to the normally-closed position. The rigidly
attached pistons are also returned to the de-energized
position.
[0008] Many different combinations of FSNC diaphragm valves and
pneumatic actuators are known in the art. Within the prior art,
well-optimized valve designs were adapted to provide minimized leak
rate when the valve is closed and adequate response when the valve
is actuated with standard pressurized air in the typical range from
40-100 psig. A well-known tradeoff exists between the need for
adequately sealed valve and a quickly closing valve, promoted by a
strongly loaded spring, and the need for fast valve opening
response. Strongly loaded springs are also notorious to promote
fast diaphragm and seat wear as well as particle generation from
the impact of the diaphragm over the valve seat.
[0009] Commonly used diaphragms are light-weight (.about.0.2 gm)
and are capable of deflection with sub-millisecond response with
the impact of relatively small forces. In contrast, pneumatic
actuators represent a substantial mass (.about.10 grams) and
additional friction (between the pistons and the cylinder) that are
burdens for high-speed actuation. Nevertheless, these mass and
friction impairments can be overcome with a combination of a
strongly energized spring (to aid in fast valve closing action) and
high-pressure actuation (to overcome the strongly energized spring
and provide fast piston acceleration). However, the necessary
tradeoff between diaphragm cycle lifetime and speed has commonly
set a limitation on pneumatically actuated FSNC valves within the
range from 25-80 msec and typically within the range from 40-50
msec. Within these performance limitations, pneumatically actuated
FSNC valves have been proven to be useful and adequate for most
applications with cycle lifetimes within the range of
1,000,000-10,000,000 cycles, which is proven to be cost effective
and appropriate.
[0010] Alternatively, diaphragm actuation was implemented with
electromechanical (electrically driven, typically solenoid driven)
actuators. In this case a valve stem is settled into the normally
closed position with a preloaded spring. The stem can be pulled
away from the seat by means of electromagnetic energy. For example
U.S. Pat. No. 6,394,415 discloses valve apparatus that is capable
of 3-5 msec open-close valve response time. While this technology
represents a speed improvement over conventional FSNC pneumatic
valves it is currently limited to significantly small conductance
(C.sub.v=0.1) and low temperature operation.
[0011] Diaphragm actuation was also implemented with piezoelectric
actuators. These actuators are relatively fast with response time
approaching the 2 msec range. While these actuators show promise
for high purity applications, they are not compatible, in their
ultrahigh-purity version, with FSNC needs. In addition, conductance
is relatively limited at the C.sub.v<0.1 range.
[0012] The prior art implemented high purity and ultrahigh purity
diaphragm valves with the metallic diaphragm serving both as the
seat sealing member and the ambient sealing member. This design
advantageously minimizes sources of contamination and fluid
entrapment as described in the prior art. However, diaphragms were
occasionally subjected to catastrophic failure such as rupturing
and cracking with subsequent potentially hazardous leakage of
dangerous and/or environmentally incompatible fluid into the
ambient. In particular, reactive or toxic gasses were occasionally
released into the ambient by failing diaphragm valve. This
impairment resulted in significant safety and environmental concern
and subsequent costly measures to minimize the hazards such as
de-rated cycle lifetimes, ventilated and tightly monitored cabinets
and multiple containment.
[0013] The art of ultrahigh purity diaphragm valves has been a late
follower of a well developed diaphragm valve technology that is
known and well-documented for over a century with widespread
applications spanning from agriculture, analytical instrumentation,
plumbing, automotive, aviation, hydraulics and fluid level control
to name only a few. Diaphragm actuation with pressurizing fluid has
been practiced for many of these applications that do not mandate
FSNC valves. In this case diaphragm chambers were formed both at
the flow side and the control side (the other side of the
diaphragm). The diaphragm was flexed into sealing position by
supplying pressurized fluid into the diaphragm control chamber.
Valve response time directly corresponds and faithfully follows the
timing of fluid pressurization (valve set to be shut-off) and
de-pressurization (valve is relieved back to the normally open
state). Many useful devices and manifolds were implemented with
fluid controlled diaphragm valves such as pressure regulators and
self-compensating shut-off valves. Fluid controlled diaphragm
valves were utilized for many applications that do not mandate FSNC
design. For example, fast gas introduction into chromatography
analytical instruments. For example, U.S. Pat. No. 4,353,243
discloses an embodiment for a direct fluid actuated diaphragm
valve, configured and suitable for sample introduction within gas
chromatography applications. Embodiments within this patent and
other patents have successfully implemented polymer or elastomer
based diaphragms for adequately performing valve seal with a simple
seat design including only a flat surface and a port. U.S. Pat. No.
4,353,243 also suggested the possible utilization of a metallic
diaphragms wherein an adequate seal might be obtained by means of a
polymer coating over the internal area of the diaphragm.
[0014] Conventional ultrahigh-purity diaphragm and valve seat
designs are mostly suitable for mechanical actuation, localized at
the center of the diaphragm, which was practiced in the prior art.
In contrast, fluid actuation, by virtue of applying a uniformly
distributed force has the tendency to spread the inverted part of
the diaphragm across an area that is substantially larger than the
common valve seat. Accordingly, conventional fluid-controlled
diaphragm valves were designed for large area contact between the
diaphragm and a flat seat. However, this design is not compatible
with high purity valves wherein large area contacts are
disadvantageous. Additionally, an area based leak-tight sealing is
not practically possible with metallic diaphragms.
[0015] Diaphragm valves are inherently limited in conductance.
Valve conductance is restricted by the limited range of diaphragm
flexing. There is a recognized tradeoff between diaphragm cycle
lifetime (the number of cycles until failure) and the increase in
diaphragm flexing (to increase conductance). Accordingly, standard
size high-purity diaphragm valves were limited in conductance to
the C.sub.v range from 0.05-0.50. Where C.sub.v represents the flow
through a valve under a standard pressure gradient of 1 psi. For
example, a C.sub.v range from 0.1-0.5 represents a valve path
opening in the approximate range from 2-16 mm.sup.2 of area. It is
well known in the art that diaphragm cycle lifetime is adversely
impacted by increased range of diaphragm flexing making higher
conductance valve, generally less reliable.
[0016] Modified diaphragms were invented for increased conductance
while minimizing the tradeoff of cycle lifetime. For example U.S.
Pat. No. 5,201,492 discloses a high purity valve embodiment wherein
the diaphragm comprises several annular surfaces that are stepped
upward from a plane in which the diaphragm perimeter is secured to
the valve body. Accordingly, larger and more consistent conductance
was materialized. In the art of gas pressure sensors, corrugated
and rippled flexible-metallic-diaphragms were used to improve the
performance and reliability of pressure sensing devices, for
example, the embodiments disclosed in U.S. Pat. No. 4,809,589.
[0017] Diaphragm valves are typically limited to operate within the
temperature range that is compatible with the valve seat material.
For example, typical ultrahigh-purity diaphragm valves were
successfully implemented with Kel-F (PCTFE) seat material. Kel-F
has been implemented with superior reliability in the temperature
range up to 65.degree. C. while maintaining a resilient and
leak-tight seal. Higher operation temperature, typically up to
125.degree. C., was attainable with the aid of polyimide polymer
seat material such as Vespel.RTM.. Adequate leak integrity over
much harder Vespel seats typically requires to strengthen the
preloaded spring. To match the opening speed to the closing speed,
high temperature valves are typically actuated at higher air
pressure in the range from 60-100 psig. Accordingly, higher
temperature valves can be actuated faster than low temperature
valves. However, the resulted higher stem impact on the diaphragm
adversely shortens the cycle lifetime of diaphragms with adverse
impact on the reliability and cleanliness. In addition, the
diaphragm slamming over a much harder seat material, such as Vespel
inevitably accelerates diaphragm and seat wear and particle
formation. Vespel is considerably more brittle than other lower
temperature seat materials such as Kel-F. While Vespel based higher
temperature valves have been offered in the commercial market for
several years they are still immature and inadequate for most
applications.
[0018] High purity and ultra-high-purity (UHP) valves were
successfully installed for reliable and cost effective
functionality of many different processing equipment such as
chemical vapor deposition (CVD), physical vapor deposition (PVD)
and etching. In these applications, valves are typically cycled
once during process. Accordingly, reliable and contamination-free
cycle lifetime in the range from 1,000,000-10,000,000, that was
tested and specified for these valves, enabled the processing of
many substrates with valve actual lifetime exceeding 5 years.
[0019] In recent years, the art of semiconductor processing and
similar arts have created a commercial market for multiple valve
manifolds. Within multiple valve manifolds, several valves are
connected into a functioning control device wherein simultaneous
and/or coordinated actuation of several valves with precision is
essential. For example, the common-functionality of routing a fluid
entering from one common port into either one of two "non-common"
ports requires the synchronized actuation of two separate valves.
In the art of multiple valve manifolds, valve state uncertainty,
during the period of valve response time is undesired. Within
larger manifolds of three valves and more gas counter-flow may
result if the valves are operated out of synchronization.
Particular applications of reactive gas mixing manifolds cannot
tolerate counter-flow and require sophisticated and functionality
impaired valve delay actuation to avoid source gas and manifold
contamination. Accordingly, conventional FSNC valves with their
associated 40-50 msec response time are inadequate for many of
these forefront applications in the semiconductor, display and
pharmaceutical manufacturing industries, to name a few.
[0020] In was further recognized, in recent years that the speed
and the synchrony of valves can be improved by integrating a pilot
valve together with a FSNC valve wherein the delay and
inconsistency associated with pneumatic hoses is avoided. For
example U.S. Pat. No. 5,850,853 describes an assembly of a
conventional FSNC pneumatic valve with a standard solenoid valve
wherein the air pressure is fed into the solenoid valve and the
valve is actuated by controlling electrical current to the pilot
valve. Unfortunately, integrated pneumatic-pilot valves do not
represent a substantially improved prior art in terms of speed and
cycle lifetime.
[0021] In recent years, atomic layer deposition (ALD), a variant of
CVD has emerged as the future work-horse deposition method for
critical thin film applications. ALD is a cyclic process carried
out by dividing conventional chemical vapor deposition (CVD)
process into an iterated sequence of self-terminating process
steps. An ALD cycle contains several (at least two) chemical dose
steps in which reactive chemicals are separately delivered into the
process chamber. Each dose step is typically followed by an inert
gas purge step that eliminates the reactive chemicals from the
process space prior to introducing the next precursor.
[0022] ALD films of practical thickness typically require between
several tens to several thousands of valve cycles per layer. In
contrast, most other processes such as CVD, PVD, etching etc. are
practiced with only one valve cycle per layer. Accordingly, much
higher standards for valve cycle lifetime are required for cost
effective ALD performance. Additionally, cost-effective ALD
mandates typical valve cycle times on the order of 10-100 msec and
acceptable valve response time must be limited to 5 msec, or less.
Moreover, efficient switching delivery of low-volatility ALD
precursors with limited volatility imposes higher specifications
for valve conductance and temperature rating than the
specifications of currently available high purity valves.
[0023] For example, an ALD process with 200 cycles wears out the
valves at least 200 times faster than a CVD process, therefore
reducing practical valve lifetime from 5 years into a mere 10 days
for a valve with cycle lifetime of 1,000,000 cycles. It was also
discovered, by the inventor of this invention and others, that
under high throughput ALD conditions wherein valves are cycled
within 10-150 msec, off-the-shelf valves are typically wearing
about 10 times faster than their specified cycle lifetime. This
undesired phenomenon was found empirically to be the general trend
independent of valve manufacturer or model. As a result, even top
performing commercialized valves are expected to last only 5-30
days under high-productivity-ALD production environment.
[0024] Independently, the 25-80 msec response of standard UHP
valves introduces an uncontrolled timing uncertainty for valve
opening and shutting on the order of 10-40 msec. This range of
uncontrolled time mismatch is comparable and longer than typical
flow resident time during ALD purge which for high throughput ALD
is preferably set below 5 msec. With chemical dose step being
slotted into a 10-100 msec range, a possible 10-40 overlap between
ALD chemical dose steps and ALD purge steps is devastating. Even
worse, during actuation the conductance of the valves is poorly
defined and generally inconsistent. As ALD manifolds are in
particular designed for fast response, they are very sensitive to
counter-flow from non-synchronous actuation of valves. Therefore,
it is necessary to maintain valve actuation times to be
substantially shorter than valve cycle time (the time it takes to
open, keep open and close the valve) and in general, as short as
possible.
[0025] Ideally, ALD should be practiced with injection type valves
notated in the art as "pulsed valves". However, prior art injection
valves are not compatible with high purity standards. Likewise,
prior art high purity valve technology is not suitable for
injection valve applications.
[0026] To summarize, the need for improved performance of
multiple-valve-manifolds created a necessity for FSNC valves with
substantially faster response and time precision. These valves must
achieve more than an order of magnitude improved speed while
maintaining-and preferably improving valve-cycle lifetime. In
particular, a substantially improved valve response and cycle
lifetime are necessary to support the transition of ALD into
mass-production. There is also a need to increase the conductance
and the temperature rating of all-metal high purity valves while
maintaining their reliability, cleanliness and long cycle lifetime.
Finally, there is also a need for high purity FSNC injection valves
with the specifications of speed and reliability stated herein.
SUMMARY OF THE INVENTION
[0027] It is the objective of the present invention to provide a
method for gas flow switching with less than several milliseconds
and preferably with sub-millisecond response while maintaining the
standards of fail-safety and purity that is customary in the
technology of semiconductor processing and the like. Improving the
cycle lifetime of high-purity and ultrahigh-purity valves is also
within the scope of the present invention. It is also the objective
of the present invention to improve the safety and environmental
protection of high purity and ultrahigh purity valves and
valve-manifolds. In additional scopes the invention provides
innovative valve and seal design that increase valve conductance
and elevated temperature performance including the usage of rippled
diaphragms, bellows and advanced elastomer and metal seals.
[0028] In some embodiments the present invention discloses a method
suitable for diaphragm mounting that effectively provides
compatibility of standard diaphragm and standard ultrahigh purity
valve seat design with fluid actuation. In further improvement the
fail-safe actuator enables external control over valve conductance.
It is also a main objective of this invention to reduce diaphragm
and valve seat wear and substantially improve the cycle lifetime of
valves. Embodiments are presented for various useful improvements
in valve seal shape, materials and properties including utilization
of elastomers and coated elastomers and advantageous embodiments
for metallic seals.
[0029] Embodiments are presented for valve integration for pulsed
delivery of chemicals through vessel walls where vessels include
showerhead gas distribution apparatus. These embodiments are highly
suitable for ALD applications wherein multiple integrated valves
with negligible dead space between valve seat and showerhead space
are highly desired.
[0030] Referring to FIG. 1a, in one aspect of the invention, a
fluid control valve comprises a valve seat 110, a flow path through
the valve seat, a metallic diaphragm 108, a normally closed
pneumatic actuator 118, a valve control chamber 114, a pneumatic
feed line 116 and a pilot valve 144 wherein the diaphragm is
dispersed between the valve seat and the valve control chamber and
the normally closed pneumatic actuator is configured to normally
close the flow path by deflecting the diaphragm to seal over the
valve seat. The pneumatic feed line is preferably connected in
serial fluidic communication with the normally closed pneumatic
actuator. The pneumatic feed line is preferably connected in serial
fluidic communication with the pilot valve and the pilot valve is
preferably connected in serial fluidic communication with the valve
control chamber. Preferably, the pilot valve is a three way
normally open valve and the control chamber is preferably
communicated with the pneumatic feed line through this pilot valve
when the pilot valve is not actuated. Preferably, the control
chamber is disconnected from the pneumatic feed line by the pilot
valve and the control chamber is communicated to a vent or
evacuation line through the pilot valve when the pilot valve is
actuated. Preferably, the pilot valve is a solenoid valve.
Preferably, the vent port of the pilot valve is evacuated to
suppress noise, enhance speed and improve safety and environmental
protection. Preferably, the diaphragm is a dome-shaped metallic
diaphragm. In a recommended aspect of the invention the diaphragm
is preferably mounted with a preset deformation directed outwards
across from the valve seat and fastened under deformation.
Preferably, the deformation is reproducibly applied by reproducibly
pressurizing the diaphragm from the side of the valve seat after
placing the diaphragm between a sealing ledge within the valve seat
and a corresponding bonnet and lightly fastening the diaphragm
between the sealing ledge and the corresponding bonnet to maintain
sufficient fluidic flow restriction to enable to reproducibly
pressurizing the diaphragm and tightly securing the diaphragm
between the sealing ledge and the corresponding bonnet under the
reproducibly pressurizing conditions. Preferably, reproducibly
pressurizing means that the pressure is applied with full range
repeatability of better than 10%. Further, reproducibly
pressurizing preferably comprises applying ultrahigh purity
nitrogen at pressure in the range from 45-150 psig. In another
aspect the pneumatic actuator preferably includes a stem
penetrating through the wall of the valve control chamber and a
sliding seal preferably dispersed between the stem and the wall of
the valve control chamber. The volume of the valve control chamber
is preferably maintained at the minimum and preferably at less than
2 cubic centimeters. Furthermore, a rippled diaphragm is preferably
implemented to preferably increase the conductance of the valve.
Additionally, the fluid control valve is preferably supplied with
pressurized fluid to actuate the normally closed pneumatic actuator
to repel the pneumatic actuator away from the diaphragm. At the
same time, the pressurized fluid is preferably connected into the
valve control chamber when the pilot valve is preferably not
actuated to preferably deflect the diaphragm to seal over the valve
seat by the pressurized fluid and the pressurized fluid is
preferably disconnected from the valve control chamber when the
pilot valve is actuated while the control chamber is preferably
vented or evacuated when the pilot valve is actuated and the
diaphragm is flexibly snapped away from the valve seat to enable
flow through the fluid control valve. In one preferred variant the
pressurized fluid is preferably supplied to the fluid feed line
from a solenoid valve bank. Accordingly, the conductance of the
pilot valve and the volume of the valve control chamber are
preferably adjusted such that the response time of the enabled the
flow through the fluid control valve is substantially similar to
the response time of the pilot valve and preferably shorter than
two milliseconds more preferably shorter than one millisecond and
most preferably shorter than half a millisecond. The pressurized
fluid is preferably connected to the valve control chamber when the
pilot valve is de-actuated and consequently the flow through the
fluid control valve is disabled. Again, well adjusted pilot valve
conductance and minimized volume valve control chamber result with
a valve shut-off that is substantially similar to the response time
of the pilot valve and preferably shorter than two milliseconds
more preferably shorter than one millisecond and most preferably
shorter than half a millisecond. When the valve is depressurized,
either intentionally or as a result of a failure, the normally
closed pneumatic actuator preferably returns into normally closed
position and disables the flow through the fluid control valve. The
pneumatic actuator is preferably repelled away from the diaphragm
to create a restricted gap that is preferably smaller than the full
extension of the diaphragm. The restricted gap is externally
adjustable by externally adjusting the travel of the pneumatic
actuator when the pneumatic actuator is actuated and the
conductance of the fluid control valve is preferably determined by
the restricted gap that preferably limits the deflection of the
diaphragm. In another preferred modification the valve seat
preferably includes a valve seal made from an elastomer.
Preferably, the elastomer is coated by a thin layer of polymer.
More preferably the seal is preferably plated with a thin layer of
metal. Also, the seal is preferably placed within a corresponding
valve seat and the seat, including the seal are plated with thin
film of metal. Preferably the fluid control valve is applied for
controlling the pulsed delivery of gas into an ALD process
apparatus.
[0031] In another aspect of the invention, a fluid control valve
comprises a valve seat, a flow path through the valve seat, a
metallic bellow, a normally closed pneumatic actuator, a valve
control chamber, a pneumatic feed line and a pilot valve wherein
the metallic bellow is dispersed between the valve seat and the
valve control chamber to seal between the valve seat and the valve
control chamber preferably by mounting a first end of the bellow
between the valve seat and the valve control chamber and enclosing
a second end of the bellow with a substantially flat disc. The
normally closed pneumatic actuator is preferably configured to
normally close the flow path by deflecting the disc on the second
end of the bellow to seal over the valve seat. Additionally, the
pneumatic feed line is preferably connected in serial fluidic
communication with the normally closed pneumatic actuator and in
serial fluidic communication with the pilot valve while the pilot
valve is preferably connected in serial fluidic communication with
the valve control chamber. Preferably, the pilot valve is a
three-way normally open valve and the control chamber preferably
communicates with the pneumatic feed line through the pilot valve
when the pilot valve is not actuated. The control chamber is
preferably disconnected from the pneumatic feed line and the
control chamber preferably communicates with a vent or evacuation
line through the pilot valve when the pilot valve is actuated.
Preferably, the pilot valve is a solenoid valve. Preferably, the
bellow is an electroformed or an hydroformed metallic bellow. A
welded metallic bellow together with a spring implemented to
maintain the welded metallic bellow at a preset compressed position
is also preferred. The pneumatic actuator preferably includes a
stem penetrating through the wall of the valve control chamber and
a sliding seal dispersed between the stem and the wall of the valve
control chamber. Preferably, the volume of the valve control
chamber is less than 2 cubic centimeters. When the fluid feed line
is preferably supplied with pressurized fluid the pressurized fluid
actuates the normally closed pneumatic actuator to repel away from
the bellow and at the same time to communicate into the valve
control chamber when the pilot valve is not actuated and preferably
deflect to seal the disc over the valve seat by the pressurized
fluid. The pressurized fluid is preferably disconnected from the
valve control chamber and the control chamber is vented when the
pilot valve is actuated and the bellow is flexibly snapped away
from the valve seat to enable flow through the fluid control valve.
Preferably, the pressurized fluid is connected to the valve control
chamber when the pilot valve is de-actuated and the flow through
the fluid control valve is disabled. The response time of the valve
is preferably minimized by preferably maintaining the conductance
of the pilot valve sufficiently high and preferably minimizing the
volume of the valve control chamber to obtain response time of that
is substantially similar to the response time of the pilot valve.
Preferably, the response time is shorter than 4 milliseconds. More
preferably, the response time is shorter than 2 milliseconds and
most preferably, the response time is shorter than 1 millisecond.
Preferably, if the fluid feed line is depressurized either
intentionally or due to failure the normally closed pneumatic
actuator preferably returns into normally closed position and the
normally closed pneumatic actuator at the normally closed position
disables the flow through the fluid control valve. Externally
adjusted conductance is preferably desired and preferably
implemented when the pneumatic actuator is repelled away from the
bellow to create a restricted gap that is smaller than the full
compression of the bellow. Preferably, the restricted gap is
externally adjustable by externally adjusting the travel of the
pneumatic actuator when the pneumatic actuator is actuated and the
conductance of the fluid control valve is determined by the
restricted gap when it is smaller than the full compression of the
bellow. In a preferred variant the valve seat includes a valve seal
wherein the seal is made from an elastomer. In a more preferred
variant the elastomer is coated by a thin layer of polymer and in a
most preferred variant the seal is plated with a thin layer of
metal. Preferably, the seal is placed within a corresponding valve
seat and the seat, including the seal, is plated with thin film of
metal. The valve is preferably used for controlling the pulsed
delivery of gas into an ALD process apparatus.
[0032] In an additional aspect the invention teaches a fluid
control valve comprising a valve body wherein an inlet and an
outlet ports are formed, a valve camber bottom portion formed in
the valve body wherein a first port is connected in serial fluidic
communication into the valve bottom portion substantially at the
center of the valve chamber bottom portion, the second port is
connected in serial fluidic communication into the valve bottom
portion substantially off the center of the valve chamber bottom
portion, a valve seal located inside the valve chamber bottom
around the first port and a valve chamber top portion made from a
substantially flexible member wherein the center of the
substantially flexible member is normally positioned substantially
separated from the valve chamber bottom portion. Further, the valve
chamber top preferably separates the valve chamber from a valve
control chamber comprising a fluid connection port and a
translatable stem. The translatable stem is preferably actuated by
pressurized fluid means through a fluid feed line and the fluid
connection port into the valve control chamber is preferably
connected in serial fluidic communication with the fluid feed line
through a pilot valve. The fluid path in the pilot control valve
normally connects the fluid from the fluid feed line to the fluid
connection port of the valve control chamber. When the pilot valve
is actuated, preferably the fluid path in the pilot valve
disconnects the fluid from the fluid feed line to the fluid
connection port of the valve control chamber and vents or evacuates
the valve control chamber. The valve stem is preferably normally
compressed with a spring to push and deflect the flexible member
between the valve chamber and the control chamber to conform and
substantially seal over the valve seal. Fluid is preferably applied
through fluid feed line to actuate the valve stem to translate away
from the flexible member between the valve chamber and the control
chamber and at the same time the fluid is preferably applied into
the valve control chamber through the pilot valve when the pilot
valve is not actuated to deflect the flexible member between the
valve chamber and the control chamber to conform and substantially
seal over the valve seal. When the pilot valve is actuated, the
fluid is preferably vented or evacuated out of the valve control
chamber and as a result the flexible member returns to a
free-standing position when the pilot valve is actuated and the
fluid control valve is open. Preferably, the pilot valve is a
solenoid valve. Preferably, the flexible member is a dome-shaped
diaphragm. The dome-shaped diaphragm is preferably reinforced at
the perimeter wherein the reinforcement preferably comprises
mounting the diaphragm under pressurizing deflection applied from
the concave side of the diaphragm. In another preferred variant the
flexible member comprises a metallic bellow. The metallic bellow is
preferably assembled with a return spring 912 that maintains the
bellow in a substantially compressed form wherein the bellow and
the return spring 912 are fastened together as a bellow-spring
assembly and the fluid control valve is open when the bellow-spring
assembly is at the free standing form.
[0033] In yet another aspect of the invention, a valve seat
assembly comprising a perimeter seal, a perimeter groove
corresponding to the perimeter seal having a substantially circular
cross section at the top and a perimeter ledge located
substantially at the bottom wherein the perimeter seal and the
perimeter groove substantially match is disclosed. Preferably, the
perimeter seal comprises a core elastomer body and a thin polymer
coating. Preferably, a thin polymer coating over the core elastomer
body is further plated with thin metallic film conformally covering
the surface of the perimeter seal.
[0034] In another aspect, a method for preparing and mounting a
valve seat is taught comprising forming a perimeter elastomer seal.
The perimeter seal preferably has a substantially circular
cross-section on the top and a perimeter mounting ledge on the
bottom wherein the method preferably includes forming a perimeter
groove in the valve seat and the perimeter groove corresponds to
the mounting ledge of the perimeter seal and further bonding the
perimeter seal to the perimeter groove.
[0035] In an additional aspect a method for preparing and mounting
a valve seat is disclosed comprising forming a perimeter elastomer
seal, the perimeter seal having a substantially circular
cross-section on the top and a perimeter mounting ledge on the
bottom, and further coating the perimeter seal with a thin layer of
polymer, forming a perimeter groove in the valve seat where the
perimeter groove corresponds to the mounting ledge of the perimeter
seal and bonding the perimeter seal to the perimeter groove.
[0036] In an additional scope, a method for preparing and mounting
a valve seat comprises forming a perimeter, preferably radial,
elastomer seal wherein the perimeter seal has a substantially
circular cross-section on the top and a perimeter mounting ledge on
the bottom. Furthermore, the method preferably includes coating the
perimeter seal with a thin layer of polymer, activating the surface
of the thin layer of polymer for electroless plating, coating the
perimeter seal with a thin layer of metal using electroless plating
or combination of electroless plating and electroplating, forming a
perimeter groove in the valve seat wherein the perimeter groove
corresponds to the mounting ledge of the perimeter seal, placing
the perimeter seal into the perimeter groove and plating the valve
seat with a thin layer of metal wherein the thin layer of metal
conforms to the surface of the perimeter seal and the valve seat.
Preferably the metal film is nickel or a nickel alloy.
[0037] In yet an additional aspect, a fluid control valve comprises
a valve seat, a flow path through the valve seat, a metallic
diaphragm, a normally closed pneumatic actuator, a valve control
chamber, a pneumatic feed line and a pilot valve is taught wherein
the diaphragm is dispersed between the valve seat and the valve
control chamber, the normally closed pneumatic actuator is
configured to normally close the diaphragm and the pneumatic feed
line is connected in serial fluidic communication with the normally
closed pneumatic actuator and the pilot valve. The pilot valve is
preferably connected in serial fluidic communication with the valve
control chamber. The control valve is preferably formed on the wall
of a gas distribution space. The valve seat preferably defines a
flow outlet from the fluid control valve and the flow outlet from
the fluid control valve is preferably substantially coplanar with
the wall of the gas distribution chamber.
[0038] In an additional scope, the invention discloses a fluid
control valve comprising a valve seat, a flow path through the
valve seat, a metallic bellow, a normally closed pneumatic
actuator, a valve control chamber, a pneumatic feed line and a
pilot valve, wherein the metallic bellow is dispersed between the
valve seat and the valve control chamber. The metallic bellow
preferably seals between the valve seat and the valve control
chamber including mounting the first end of the bellow between the
valve seat and the valve control chamber and enclosing the second
end of the bellow with a substantially flat disc. The normally
closed pneumatic actuator is preferably configured to normally
close the fluid control valve by deflecting the disc on the second
end of the bellow to seal over the valve seat. The pneumatic feed
line is preferably connected in serial fluidic communication with
the normally closed pneumatic actuator and the pilot valve and the
pilot valve is preferably connected in serial fluidic communication
with the valve control chamber. The control valve is preferably
formed on the wall of a gas distribution space. The valve seat
preferably defines a flow outlet from the fluid control valve and
the flow outlet from the fluid control valve is preferably
substantially coplanar with the wall of the gas distribution
chamber.
[0039] The invention provides apparatus and method for fail-safe
normally closed (FSNC) pneumatically actuated valves with less than
2 msec time response, and preferably with high conductance up to
C.sub.v of 5 and preferably high temperature operation in excess of
300.degree. C. Preferably, the FSNC pneumatic valves according to
the invention withstand more than 10 million cycles wherein the
cycle time is shorter than 100 milliseconds and more preferably the
FSNC pneumatic valves according to the invention withstand more
than 50 million cycles wherein the cycle time is shorter than 100
milliseconds.
[0040] In another scope of the invention, a method for preparing
and mounting a valve seat comprises electroforming a perimeter
seal. The perimeter seal preferably has a substantially circular
cross-section on the top and a perimeter opening at the bottom.
Further the method preferably includes forming a perimeter groove
in the valve seat wherein the perimeter groove has a perimeter
corner corresponding to the perimeter opening at the bottom of the
perimeter seal, placing the perimeter seal into the perimeter
groove, brazing the perimeter seal into the perimeter groove and
plating the valve seat with a thin layer of metal wherein the thin
layer of metal conforms to the surface of the perimeter seal and
the valve seat. Preferably, the pressure of an ambient gas within a
brazing furnace is controlled and a pre-set pressure of gas is
entrapped within the perimeter seal to determine the resilience of
the seal.
[0041] In another aspect of the invention, a method for preparing
and mounting a valve seat comprises electroforming a perimeter seal
having a substantially circular cross-section on the top, a
perimeter opening at the bottom and perimeter ledges appropriately
shaped for electron beam welding, forming a perimeter groove in the
valve seat wherein the perimeter groove has a perimeter corner
corresponding to the perimeter opening at the bottom of the
perimeter seal and the perimeter corner is preferably appropriately
shaped for electron beam welding, placing the perimeter seal into
the perimeter groove, welding the perimeter seal into the perimeter
groove preferably using electron beam welding or similar means and
plating the valve seat with a thin layer of metal wherein the thin
layer of metal conforms to the surface of the perimeter seal and
the valve seat.
[0042] In yet another scope of the invention, a fluid control valve
is disclosed comprising a valve port, a mechanical valve actuator
movable between a closed valve position and a open valve position,
a valve actuator driver for driving the mechanical valve actuator
from the closed valve position to the open valve position and a
pneumatic valve driver for pneumatically opening and closing the
valve port when the mechanical valve actuator is in the open valve
position. Preferably, the valve further includes a valve diaphragm
located between the valve port and the mechanical valve driver. The
pneumatic valve driver preferably acts directly on the valve
diaphragm. The valve actuator driver preferably comprises a
pneumatic actuator. The mechanical valve actuator preferably
comprises a spring.
[0043] In an additional aspect of the invention, a fluid control
valve comprises a valve port, a mechanical valve actuator movable
between a valve closed position and an active valve position, a
valve actuator driver for driving the mechanical valve actuator
from the closed valve position to the open valve position and a
pneumatic valve driver for pneumatically opening and closing the
valve port when the mechanical valve actuator is in the active
valve position.
[0044] In one aspect, a method of operating a fluid control valve
comprises mechanically holding the valve closed in an inactive
state in which it cannot be operated pneumatically is taught and
further includes changing the valve to an active state in which it
can be opened and closed pneumatically, and opening and closing the
valve pneumatically. Preferably, the changing comprises
pneumatically actuating a mechanical valve actuator. Mechanically
holding preferably comprises holding the valve closed with a
spring.
[0045] In an additional variant, a method of operating a fluid
control valve comprising holding the valve diaphragm closed with a
mechanical actuator, releasing the mechanical actuator and opening
and closing the valve diaphragm pneumatically is taught.
Preferably, the releasing is performed pneumatically. Simultaneous
with the releasing, pneumatic pressure is preferably substituted
for mechanical pressure to hold the valve closed.
[0046] In another advantageous aspect of the invention, a fluid
control valve is disclosed comprising a valve seat, a flow path
through the valve seat, a flexible member 62, a pneumatic actuator
64, a flexible member chamber 54, a flexible member chamber
evacuation port 56 and an evacuation line 70, wherein the flexible
member is dispersed between the valve seat and the flexible member
chamber. The pneumatic actuator is preferably configured to close
the flow path by deflecting the flexible member to seal over the
valve seat. The flexible member chamber is preferably pressure
sealed and the flow path preferably remains pressure sealed from
the ambient when a flexible member failure occurs. Preferably, the
flexible member comprising a metallic diaphragm or a metallic
bellow. Preferably, the flexible member chamber is preferably
further evacuated following a failure of the flexible member.
[0047] This invention includes multiple improvements that
individually or in combinations provide a substantial advancement
of fail-safe normally close valve technology. Embodiments within
the invention provide solutions for many of the well-known
deficiencies in the prior art. Actual utilization of various
improvements depends on the specific application. The preferred
embodiments that are described below serve to present and clarify
the apparatus, method and improvements of all the various aspects
of this invention. Those who are skilled in the art can use the
details given below to select the proper and cost-effective
combination that is most suitable for a given application.
BRIEF DESCRIPTION OF THE DRAWINGS
[0048] The accompanying drawings, which are incorporated in and
form a part of the specifications, illustrate the preferred
embodiment of the present invention, and together with the
description serve to explain the principles of the invention. In
the drawings:
[0049] FIG. 1 depicts an illustrative cross-sectional side view of
a high-speed FSNC diaphragm valve in accordance with the invention.
A pneumatic manifold is depicted in flow schematic form for
improved clarity. The valve is depicted in the "active" mode and
the "shut" state in FIG. 1a. The valve is depicted in the "active"
mode and the "open" state in FIG. 1b. The valve is depicted
illustrating the pilot valve manifold in FIG. 1c. The time response
of the valve is depicted in FIG. 1d. A generalized implementation
of a safety enhanced valve is depicted in FIG. 1e.
[0050] FIG. 2 depicts an illustrative cross-sectional side view of
a high-speed FSNC diaphragm valve in accordance with the invention
wherein the pilot valve is integrated. The three states of the
valves namely "Inactive", "Active-Shut" and "Active-Open" are
depicted in FIGS. 2a, 2b and 2c, respectively.
[0051] FIG. 3 depicts illustratively the diaphragm space within an
ultrahigh-purity valve shown in the states of "open" and "shut" and
in transition states between "shut" and "open".
[0052] FIG. 4 depicts an illustrative side cross-sectional view of
high conductivity valve utilizing a rippled diaphragm in accordance
with the current invention. The valve is illustrated in the "active
shut" (top) and the "active open" (bottom) states.
[0053] FIG. 5 depicts an illustrative side cross-sectional view of
a conventional diaphragm valve (top) and a high conductivity valve
utilizing a rippled diaphragm (bottom) in accordance with the
current invention. The valves are illustrated in the "active open"
state.
[0054] FIG. 6 depicts an illustrative cross-sectional view of a
FSNC high-conductivity valve based on a formed bellow in accordance
with the current invention.
[0055] FIG. 7 depicts an illustrative side cross-sectional view of
a FSNC high-conductivity valve based on a formed bellow in
accordance with the current invention. The valve is shown in the
"active shut" (top) and the "active open" (bottom) states.
[0056] FIG. 8 depicts an illustrative side cross-sectional view of
a FSNC high-conductivity pulsed valve based on a formed bellow in
accordance with the current invention. The valve is shown in the
"active shut" state.
[0057] FIG. 9 depicts an illustrative side cross-sectional view of
a FSNC high-conductivity pulsed valve based on a formed bellow in
accordance with the current invention. The valve is shown in the
"active open" state (top) and the "active shut" state (bottom).
[0058] FIG. 10 depicts a side cross-sectional view of a round seal,
implemented within a high-purity valve in accordance with the
current invention. A round elastomer based o-ring seal is firmly
mounted with a ledge (FIG. 10a). An elastomer coated with a polymer
film is depicted schematically in FIG. 10b.
[0059] FIG. 11 depicts a metal-coated elastomer-seal in accordance
with the current invention showing the process flow for an
integrated metallic seal.
[0060] FIG. 12 depicts the process flow for making and integrating
a metal seal in accordance with the current invention.
[0061] FIG. 13 depicts the process flow for making and integrating
a metal seal in accordance with the current invention.
[0062] FIG. 14 depicts several examples of integrated metal seals
in accordance with the current invention.
[0063] FIG. 15 depicts a schematic side cross-sectional view of a
high conductivity FSNC bellow pulsed-valve in accordance with the
invention.
[0064] FIG. 16 depicts a schematic side cross-sectional view of a
high conductivity FSNC bellow pulsed-valve in accordance with the
invention shown in the "active open" state (top) and the "active
shut" state (bottom).
[0065] FIG. 17 depicts a schematic side cross-sectional view of a
high conductivity FSNC bellow pulsed-valve, suitable for very
high-temperature applications, in accordance with the invention
shown in the "active shut" state.
[0066] FIG. 18 depicts a side cross-sectional view of an ALD
manifold comprising four pulsed-valves and a showerhead in
accordance with the current invention.
[0067] FIG. 19 depicts a side cross-sectional view of a fluid
actuated diaphragm illustrating a standard conventionally and
commonly mounted dome-shaped diaphragm (19a), a pre-loaded
diaphragm suitable for optimized fluid actuation (19b), an improved
seat suitable for optimized fluid control (19c), and an exemplary
rippled diaphragm that is more suitable for optimized fluid
control.
[0068] FIG. 20 depicts a side cross-sectional view of a fluid
actuated diaphragm valve illustrating the procedure of mounting a
diaphragm at small torque (FIG. 20a) followed by a diaphragm
deflection and tightening under pressure applied from the valve
seat side (FIG. 20b) to stress the perimeter of the diaphragm.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
A. Fail-Safe Pneumatic Valve with Fast Response and Extended Cycle
Lifetime
[0069] Embodiments in accordance with this invention utilize
standard and modified ultra-high purity valve seat design with a
standard diaphragm to obtain fast FSNC valves suitable for most
challenging applications. Additional embodiments use rippled
diaphragms and suitable bellow arrangements to obtain fast FSNC
valves with enhanced properties such as conductance, service
temperature and reduced dead-space.
[0070] Central to the invention is the integration of FSNC valve
actuator with a diaphragm or bellow and corresponding valve seat
that are made suitable and optimized for both stem and fluid
actuation. Conventional diaphragm and valve seat designs are mostly
suitable for mechanical actuation, localized at the center of the
diaphragm, as practiced in the prior art. In contrast, fluid
actuation, by virtue of applying a uniformly distributed force has
the tendency to spread the inverted part of the diaphragm across a
substantial area. Accordingly, as described in the introduction,
conventional fluid-controlled diaphragm valves were designed for
large area contact between the diaphragm and a flat seat. The
implementation of a useful embodiment in accordance with this
invention mandates that diaphragm and seat design are optimized for
both fluid and stem actuation. This optimization is achieved by one
or a combination of methods, described below, in reference to FIG.
19.
[0071] An ultrahigh-purity Fail-Safe-Normally-Closed (FSNC) valve
according to the invention is exemplified in FIG. 1a. Valve 100
comprises a metallic valve body 102 wherein at least one inlet
passage 104 and one outlet passage 106, are formed. Inlet passage
104 is linked into diaphragm chamber 107 through the center 110' as
commonly practiced in the art, while outlet passage 106 is linked
to diaphragm chamber 107 offset from the center as practiced in the
art. Diaphragm 108 made of suitable metallic alloy such as Elgiloy,
Hastelloy, ST40 titanium, NW4400, Inconel 625, Nimonic 115 and the
like is mounted at the perimeter 109 to seal over diaphragm chamber
107 by virtue of pressure applied by bonnet 112 and nut 111.
Diaphragm 108 is set to seal over valve-seat 110 when the valve is
commanded to the "SHUT" position. Valve seat 110 is integrated into
valve body 102 to surround the inlet of passage 104 as practiced in
the art or according to additional embodiments that are disclosed
in this invention and are described in details, below.
[0072] In the art of fluid control valves the term "valve-seat" is
utilized to define the general place wherein the valve seal is
formed. In this respect valve seat 110 commonly represents the port
110' wherein the seal is performed, the seal 110'' and the
corresponding seal mounting grooves 110'''.
[0073] Above diaphragm 108, fluid control chamber 114 is formed
between diaphragm 108, bonnet 112 and dynamic seal 122. Control
chamber 114 is equipped with fluid control port 116. Valve stem 118
is mounted through an opening 119 in bonnet 112. Translation of
stem 118 is used to actuate diaphragm 108 by the pressure of pad
120. Stem 118 and pad 120 are positioned to normally press on
diaphragm 108 by the force of spring 128. When the diaphragm is
held by stem 118 and pad 120 to shut-off the passage from inlet 104
into diaphragm chamber 107, the valve is in the "INACTIVE" state.
Pneumatic actuator 124 includes piston 126 and sliding seal
127.
[0074] Pressurized air or inert gas is supplied from source 140
through 3 way pilot valve 142 as practiced in the art. Valve 100 is
normally at the fail-safe mode. To actuate the valve from fail-safe
mode to "ACTIVE" mode the valve system is pressurized by commanding
pilot valve 142 to pressurize the valve system through supply line
146. When pneumatic actuator 124 is appropriately pressurized
through port 130, piston 126 is translated away from the valve
pulling stem 118 and pad 120 away from diaphragm 108 to command the
fail-safe mechanism into the "ACTIVE" position. Concurrently, the
pressure is supplied into control chamber 114 through normally open
pilot valve 144, integrated line 154 and port 116. Accordingly,
diaphragm 108 is deflected by the pressure into an "ACTIVE SHUT"
position. Valve system 100 is maintained in "ACTIVE" state for as
long as air pressure is appropriately supplied through line 146. In
the event that the air supply through line 146 is interrupted,
either intentionally by commanding pilot valve 142 from "PRESSURE"
to "VENT" or due to failure of the pneumatic system, the valve is
returned to the "INACTIVE" state by the FSNC action of spring
128.
[0075] When valve 100 is in the "ACTIVE" state, pilot valve 144 is
used to control the pressure within control chamber 114 from "HIGH"
to "LOW" and correspondingly the functionality of valve 100 from
"ACTIVE SHUT" to "ACTIVE OPEN" respectively. "ACTIVE OPEN" state is
commanded by setting pilot valve 144 from its "normally open" state
into a "vent" state. At "vent" state, pilot valve 144 shuts the
path to air supply line 152 and vents control chamber 114 into vent
line 156. FIG. 1b depicts valve 100 in the "ACTIVE OPEN" state.
Pneumatic manifold 160 is preferably integrated into the
self-contained assembly of valve 100. For example, assembly 160'
depicted in FIG. 1c.
[0076] FIG. 1b illustrates mechanism 170 that is useful for
externally adjusted conductance. Accordingly, the translation of
stem 118 is adjustably determined by positioning screw 172. As a
result, button pad 120 can be positioned to restrict the
back-deflection of diaphragm 108. Leak integrity of pneumatic
actuator 124 is maintained by dynamic seal 174. The conductance of
valve 100 is determined by the gap between diaphragm 108 and seal
110''. This gap is controlled by the position of button 120 when
the valve is in "ACTIVE" mode.
[0077] In FIG. 1c valve 200 is illustrated with an integrated air
manifold 160' rigidly assembled. Air is supplied through pneumatic
line 146'. Vent port 156' of pilot valve 144' is preferably
separately vented or evacuated. The air supplied through line 146'
splits at 148' into actuator line 150' and pilot supply line 152'.
Solenoid valve 144' is held "normally open" by spring 282. The vent
outlet 290 is sealed by seal 288 of puppet 280. Air is therefore
connected from inlet 286 into supply line 154' to feed inlet 116'
and command the valve to the "ACTIVE SHUT" state (depicted in FIG.
1c). When solenoid valve 144' is commanded "vent" by energizing
coil 292, puppet 280 moves against spring 282 to seal over inlet
port 286 with seal 284 and block the path to the pressurized air
supply 286. Concurrently, the vent port 290 is cleared and the air
is vented or evacuated out of control chamber 114 to set valve 200
into "ACTIVE OPEN" state.
[0078] The embodiments depicted in FIGS. 1a, 1b and 1c exemplify a
preferred embodiment of ultrahigh-purity FSNC valve system. Those
who are skilled in the art can apply this design in a variety of
useful variants that are optimized to a variety of specifications
for size, operating temperature, valve-cycle timing and duty cycle.
Embodiments must ensure that performance is not impaired due to
inappropriate air supply and/or inadequate manifold 160 design. In
particular, the air being drawn from manifold 160' to actuate valve
100 from "ACTIVE OPEN" to "ACTIVE SHUT" should not cause a
meaningful drop of air pressure at actuator 124. Also, the
conductance of pilot valve 144 (144') should be sufficient to
pressurize and to vent control chamber 114 within a specified time
to facilitate fast valve cycling. Likewise, the effective volume
for pressurizing and depressurizing through valve 144 should be
kept to the minimum to support fast valve cycling. This volume
includes control camber 114, inlet 116 and supply line 154.
Finally, both supply line 154 and inlet 116 should be designed to
maximize the conductance into control chamber 114. If these
guidelines are not observed, valve functionality is likely to be
impaired and inconsistent.
[0079] Preferably, exhaust 156' (FIG. 1c) is evacuated into a
vacuum pump to accelerate the de-pressurizing of control chamber
114 with advantageously faster valve opening speed. The speed
advantage of evacuating exhaust port 156' is demonstrated in FIG.
1d wherein the response of a fast pneumatic valve (FPV) according
to the invention is illustrated for a FPV configuration with
combined volume of diaphragm control chamber 114 and line 116' of 2
cc and a pilot valve 144' conductance of 100 P cc/sec (P, the
actuating fluid pressure, is given in psi (This conductance can be
expressed as .about.0.002 P liter/sec where P is given in Torr).
The 1.25 msec response time of pilot valve 144' is deconvoluted to
yield the "pure" FPV response. Accordingly, the pressure within
diaphragm control chamber, P.sub.DCC is ploted versus time. With 60
psi pressure the specified FPV shuts-off within less than 50
.mu.sec (active-open [AO] to active-shut [AS]). However, the
opening time (AS to AO) without evacuating port 156' extends to
.about.1 msec, more than 20 times longer than the shut-off time. In
contrast, when the exhaust port 156' is evacuated, the AS to AO
time is reduced to .about.50 .mu.sec which is comparable to the AO
to AS response time. Additionally, exhaust 156' evacuation
significantly muffles the loud audible noise that results from
high-speed exhaustion of gas from control chamber 114 into the
ambient. This noise becomes very pronounced when FPVs are actuated
with .about.1 msec speed. For example the audible noise generated
by a <2 msec responding FPV with .about.2 cc 114 volume exceeded
100 dB at the vicinity of the an ALD system. Within
high-productivity ALD equipment with typically 5-10 valves per ALD
chamber, actuating at 4-10 Hz (the purging valve are cycled twice
per ALD cycle), this audible noise amounts for a significant hazard
and inconvenience. Vent port evacuation was implemented to reduce
the noise of multiple FPVs within a high-productivity ALD manifold
to a very low level of <50 dB while at the same time maintained
a valve response time at <1.25 msec for both open and close and
enhanced the safety of chemical delivery as further described
below.
[0080] Evacuating the vent port 156' of FPV 200 substantially
improves the safety of valves and valve manifolds with containment
of possible leaks into the ambient when diaphragms rapture. As
described in the introduction the prior art high-purity and
ultra-high-purity valves present substantial safety and
environmental hazard when diaphragms rapture with subsequent
leaking of hazardous and environmentally incompatible chemicals
into the ambient. In contrast, according to the invention, FPVs are
preferably evacuated through suitable stainless steel conduits into
an appropriate vacuum pump. Preferably, the FPVs are actuated with
inert clean gas such as nitrogen. A ruptured diaphragm results in
the flow of actuation gas, i.e. nitrogen, through the ruptured
diaphragm into the process chamber and the manifold. This pressure
increase is preferably interlocked to shut down the process, shut
down chemical source valves and shut down the supply of actuation
gas to the FPV (i.e. by de-actuating valve 142 in FIG. 1a) while
concurrently evacuating the FPVs by actuating pilot valves 144'.
Accordingly, uncontrolled chemical release into the ambient is
prevented. Following the interlock action the failed FPV is
identified by individually activating a FPV followed by pilot valve
144' actuation. A faulty FPV is identified when a pilot 144'
actuation correlates with a chamber pressure increase.
[0081] This safety feature is further implemented for enhancing the
safety of stem actuated pneumatic valves wherein the advantageous
chamber 114 is implemented for the sole purpose of enhanced safety.
In this preferred embodiment the diaphragm chamber 114 is evacuated
through port 116' and preferably through a normally closed 2 way
valve in serial fluidic communication between port 116' and the
vacuum pump. Again, diaphragm rupture is handled by the interlocks
as described above, and hazardous leaks into the ambient are
prevented. This embodiment is further illustrated in FIG. 1e
wherein diaphragm 62 is mounted within diaphragm chamber 54 and the
stem 58 is dynamically sealed using seal 55. Actuator 64 is driven
by compressed gas as commonly practiced in the art to drive stem 58
away from diaphragm 62. Accordingly, diaphragm 62 is able to flex
and open the flow path (not shown). When actuator 64 is
de-pressurized the valve is returned to the normally closed
position by the action of spring 59. Diaphragm chamber 54 is linked
with vacuum pump 74 through port 56, 2 way valve 60, conduit 70,
optional abatement module 80 and line 71. In the event of
catastrophic diaphragm failure, the supply of gas to actuator 64 is
interrupted and the valve shuts-off. The chemical within the valve
penetrates into the volume of diaphragm chamber 54 which was
preferably evacuated prior to the failure event. Preferably, the
volume of diaphragm chamber 54 is evacuated and the leaking of
hazardous chemical into the ambient is prevented. In some preferred
embodiments, abatement module 80 implements means to substantially
abate the hazardous chemicals from the gas upstream from pump 74.
For example, module 80 includes an abatement surface with
maintained temperature exceeding 800.degree. C. Pump 74 is further
exhausted through atmospheric pressure conduit 76 into abatement
module 72 as commonly practiced in the art. Abatement module 72 is
capable of removing hazardous chemicals from the effluent gas and
the scrubbed gas is then emitted into the ambient using conduit 78.
Many different device processing techniques such as ALD, CVD ion
implantation and epitaxial growth reactors implement extremely
hazardous chemicals such as trimethylaluminum, arsine, phosphine,
hydrazine, tungsten hexafluoride, germane, silane etc. wherein
contained diaphragm valves (CDV per FIG. 1e) are suitable to
substantially enhance the safety of personal and the
environment.
[0082] Following the principles that are disclosed above, an
additional embodiment in accordance with the invention integrates
the pilot valve in serial fluidic communication between the
actuator (124 FIG. 1c) and the valve control chamber (114 FIG. 1c).
This embodiment is illustrated in FIGS. 2a, 2b and 2c. In FIG. 2a,
valve 300 is illustrated in the "INACTIVE" state. Accordingly, port
130' is vented and the normally closed mechanism comprising of stem
118', spring 128' and pad 120' exerts pressure on diaphragm 108 and
the flow path between ports 104 and 106 is shut. Valve 300 includes
a top fluid feeding port 130', a fluid conduit composed of gap 320,
and path 326 to direct fluid into or out-of actuator chamber 328.
Dynamic seal 322 enables the motion of stem 118' while maintaining
the integrity of fluid path 130', 320 and 326. Fluid is
communicated into the normally open pilot valve 144'' through
conduit 152'' which can be reduced into a hole in the wall of
actuator chamber 328 as shown in FIG. 2a. Some parts of pilot valve
144'' are not labeled to simplify the illustration. However, those
who are skilled in the art can draw a similarity between pilot
valve 144' in FIG. 1c and pilot valve 144'' in FIG. 2a to fully
understand the attributes of the various parts. In the normally
open position the plunger 280' of pilot valve 144'' enables serial
fluidic communication between 152'' and 116'' to deliver fluid into
valve control chamber 114' while preventing serial communication
(and venting) into vent/evacuation port 156''.
[0083] As illustrated in FIG. 2b, when fluid, such as compressed
air, 332, is introduced into port, 130', the fluid is inserted
through 326 into actuator chamber 328 to exert force on plunger
126'. As a result, plunger 126', stem 118' and pad 120' are pushed
away from diaphragm 108. Concurrently, fluid 332 is communicated
through port 152'', valve 144'' and port 116'' into control chamber
114' to exert force over diaphragm 108 and maintain the valve shut.
Accordingly, the valve is in "ACTIVE SHUT" state. Failure may
undesirably remove the pressurized fluid from port 130' or
alternatively reduce the pressure of fluid 332 below the adequate
level. In that case the normally closed mechanism comprising spring
128', stem 118' and pad 120' will return the valve to a safe,
normally closed, "INACTIVE" state.
[0084] As illustrated in FIG. 2c, when valve 300 is "ACTIVE" by
virtue of supplying adequately pressurized fluid 332 into port
130', the valve can be actuated from "ACTIVE SHUT" (FIG. 2b) to
"ACTIVE OPEN" (FIG. 2c). To open valve 300 pilot valve 144'' is
actuated. For example solenoid valve 144'' is actuated electrically
through connection 330 to pull plunger 280' against spring 282' to
seal over fluid inlet 286' and vent control chamber 114' through
vent port 156''. Inlet port 286' is sealed with sealing member 284'
while sealing member 288' is removed from vent/evacuation port
156''. As a result, diaphragm 108 can flex open by the diaphragm's
own spring force to connect port 104 with port 106.
[0085] Typically, control chambers 114 with 1-1.5 cm.sup.3 were
easily constructed for valves with standard diaphragm diameter of
.about.2.5 cm (.about.1 inch). Additional volume from port 116 and
supply line 154 typically increased the actual volume of the
diaphragm control space to 1.5-2.0 cm.sup.3. Standard 3 way pilot
valves were integrated into valve assembly 200 with conductance
typically limited into C.about.0.002 P liter/sec wherein P is the
air inlet pressure in Torr units (or .about.100 P cc/sec where P is
expressed in psi). As detailed above, in reference to FIG. 1d, the
internal time response of a FPV follows in most cases the response
time of the pilot valves which is typically substantially longer
than the characteristic gas dynamic response of a well design FPV
and the response of standard diaphragms that are shorter than 50
.mu.sec. Accordingly, FPVs can be cycled within a characteristic
time of .about.1 msec with commercially available high-speed pilot
valves. Practically, most pilot valves are limited to internal
response time that is longer than 1 msec making the valve cycle
time a replicate of the pilot valve performance. For example, valve
200 was implemented into an ALD manifold and a series 9 valve
available from Parker Hannifin General Valve division was utilized
as a pilot valve with typical response time of 1-2 msec.
Consequently, the response of the pneumatic ALD valves was
.about.1.25 msec indicating that in the time scale of .about.1
msec, the response of valve 200 is indeed determined by the pilot
valve.
[0086] A supply line 146, typically 2 meters long with internal
diameter of 4.5 mm was applied to supply air to valve 200. To
prevent significant pressure modulations and associated impaired
valve performance, when the FSNC valve is actuated between "ACTIVE
CLOSED" and "ACTIVE OPEN", the conductance of line 146 must be
substantially larger than the conductance of pilot valve 144.
Indeed the conductance of the 0.25'' OD pneumatic line (with a 4.5
mm ID), in one preferred embodiment, was C.sub.c.about.0.037
P(Torr) or 18.5 times larger than the conductance of pilot valve
144 meaning that the pneumatic pressure at tee 148 (FIG. 1a) was
substantially and adequately maintained.
[0087] The cycle lifetime of valve 200 is determined by the wear of
both diaphragm 108 and seat 110. Dome shaped diaphragm 108 is
mounted at the perimeter 109. Therefore, diaphragm 108 initially
flexes by inverting the center of the diaphragm. FIG. 3 depicts a
close-up look at diaphragm 108. The diaphragm at the stress free
state 108' is dome-shaped. When the diaphragm is stressed down
(108''), either by stem 118 and pad 120 or by a pressurized control
chamber 114, it develops inversion 302 at the center and an annular
ripple 304. When the diaphragm is stressed further (108'''), both
inversion area and ripple propagate outwards while the inverted
center of the diaphragm flexes deeper. Finally, the diaphragm makes
contact with valve seat 110. The seat prevents the diaphragm from
moving deeper. At this point the diaphragm settles into a balanced
equilibrium 108'''' wherein ripple 304 extends slightly to exert
additional loading force on seat 110. Fluid actuation between
"ACTIVE SHUT" and "ACTIVE OPEN" allows the diaphragm to flex
following its natural tendency, as depicted in FIG. 3. Accordingly,
diaphragm cycle lifetime is extended.
[0088] Due to their minimal mass, diaphragms can be flexed from a
relaxed, normally open, position to a deformed closed position
within less than 1 msec with relatively small exerted-force. For
example, the central part of a standard diaphragm weighting only
.about.0.1 gm can be moved a typical 1 mm between open and shut
position within 0.1 msec by the effect of only .about.2 Kgm. force.
This force can be easily applied with .about.1 Atmosphere (Atm.) of
air pressure. More typically, the introduction of fluid pressure,
determined by the speed of a pilot valve will determine the speed
of diaphragm actuation. For example, commercially available pilot
valves with 2 msec response can be used with diaphragm actuation
following the introduction of fluid with only a negligible
lag-behind, as explained above, with reference to FIG. 1d.
[0089] Advantageously, a typical diaphragm does not require more
than 0.5-1 Atm. of pressure to reach the seat. Additional 1-4 Atm.
is necessary to ensure seal integrity. However, the fast responding
diaphragm is positioned in contact with the seal soon before the
diaphragm chamber is fully pressurized. Once the diaphragm 108''''
makes contact with the seat, additional force is converted into
loading the diaphragm spring action through to motion of ripple 304
outwards. Accordingly, diaphragm and seat damage related to the
impact of the diaphragm on the valve seat is minimized by virtue of
the very small momentum of the moving diaphragm and the conversion
of kinetic energy into a stored loading energy by the motion of
ripple 304 outwards.
[0090] In contrast, a typical valve stem-piston assembly weighs
.about.10 grams. Accordingly, stem 118 slams diaphragm 108 on
valve-seat 110 with significant impact due to significant momentum
transfer. Additionally, kinetic energy dissipation may result with
local heating at the minimized contact area between the diaphragm
and the seal, and additional diaphragm and valve-seat wear. For
example, a 40 msec responding stem can accelerate to an estimated 5
cm/sec speed. Momentum transfer into diaphragm 108 and valve seat
110 can be 5.times.10.sup.-4 m.times.Kgm. Therefore, each cycle
dumps .about.12.5 .mu.Joules of kinetic energy into diaphragm 108
and seat 110. While neither the impact nor the energy dissipation
per valve cycle are large, their accumulated effect over tens of
thousands of cycles, and more, results in an eventual valve
failure. In particular, the indications that the valve
cycle-lifetime decreases .times.10 below specifications when the
cycle time drops below .about.1 sec suggest that the damage might
be related to accumulative stress, mechanical or thermal or both,
within diaphragm 108, valve seat 110, or both.
[0091] A method of extending the cycle lifetime of valve 200 in
accordance with the current invention preferably maintains the
valve actuated into the "ACTIVE" state where it is being further
actuated between "ACTIVE SHUT" and "ACTIVE OPEN" by pilot valve
144. For example, an ALD process may include a single cycle of
"ACTIVE"-"INACTIVE" per substrate processing wherein each valve 200
is activated at the beginning of a film run and deactivated at the
end of a film run. In another example a manifold comprising valves
200 utilized to run a CVD process may be activated whenever the
deposition chamber is initialized into a standby mode and is ready
to process wafers, and deactivated only when the chamber is
commanded out of standby mode and into a service mode. Limiting the
majority of valve cycling to fluid-actuation, and minimizing the
utilization of the damaging stem-actuated cycles, therefore
substantially extending the cycle lifetime of the diaphragm and
seat making them suitable for cost-effective ALD.
B. High Conductance Diaphragm Valves:
[0092] Cycle lifetime of diaphragm valves correlates with the
displacement of the diaphragm. Substantially minimized diaphragm
and seat damage is achieved with valves 200 apparatus and method,
as described in the previous section. Accordingly, the improved
reliability associated with fluid actuation can be, to some extent,
traded-off allowing higher conductance valves which are
advantageous for ALD. However, as explained above, the extended
cycle lifetimes specifications that are necessary for
cost-effective ALD do not leave substantial room for tradeoff.
[0093] In another embodiment of the present invention rippled
diaphragms are implemented to further extend the attainable valve
conductance within a specified cycle lifetime. For example FIG. 4
depicts valve 400 with a rippled diaphragm 408 in the "ACTIVE SHUT"
(top) and "ACTIVE OPEN" (bottom) states. To simplify the
illustration the valve section above the diaphragm seat includes
only bonnet 112, control chamber 114 and port 116. However, it
should be understood that valve 400 should be preferably
implemented and operated in accordance with the specifications and
instructions given above in reference to valve 100 (FIG. 1a), valve
200 (FIG. 1c), valve 50 (FIG. 1e) and valve 300 (FIG. 2a) and their
equivalents. Rippled diaphragms are manufactured according to
specific designs to include a perimeter 409 and several annular
preformed ripples 405. These diaphragms can be ordered at any given
design, for example, from Bellow Kuze Co., LTD. Compared to
dome-shaped diaphragms, rippled diaphragms achieve higher and more
linear spring constants with thinner diaphragms. Rippled shaped
diaphragms can be also produced by electroforming multiple layers
of plated nickel alloys and other useful films over a pre-shaped
mandrel as known in the art of electroforming and described further
below in reference to the creation of metallic seals. With improved
linearity of spring constants and the ability to employ thinner and
multiple layered diaphragms, rippled diaphragms accommodate longer
travel that translates into higher conductance valves. FIG. 5
depicts a close-up look into the difference between diaphragm
travel attainable with dome-shaped diaphragm 108 (top) and rippled
diaphragm 408 (bottom). Both diaphragms are shown in both the
"ACTIVE SHUT" (dashed) and "ACTIVE OPEN" (solid) states. The
ripple-shaped diaphragm is compatible with the specifications for
ultra-high-purity valves.
[0094] Even higher conductance can be achieved by replacing the
diaphragm with Electroformed or hydroformed metal bellows. FIG. 6
depicts an embodiment of a FSNC pneumatic valve implementing
hydroformed (or electroformed) bellow assembly 508 to serve as the
translatable sealing member within valve 500. The main design
features of valve 500 correspond to valve 100 (FIG. 1a) and valve
200 (FIG. 1c) as described before. Bellow 508 is terminated with an
open disk 505 on one side, and a closed disk 501 on the other side.
The bellow is mounted at the perimeter 509 by the pressure of
bonnet 512 to create a valve chamber 507 and a control chamber 514.
The fail-safe mechanism implements stem 518 and pad 520, pneumatic
actuator 524 and other components are generally similar to the
design of valve 100 and valve 200 that was described before. Port
516 serves to introduce fluid from a pilot valve (not shown) to
provide the fast actuation between "ACTIVE SHUT" and "ACTIVE OPEN"
states. Hydroformed bellows such as 508 are able to integrate into
the construction of a FSNC valve while maintaining the ultrahigh
purity performance. Typically, hydroformed bellows do not require
additional return spring. The travel of bellow 508 can be extended
by adding more convolutions 503 while maintaining a minimized
volume for chamber 514 to enable fast valve response. FIG. 7
depicts a schematic representation of valve 500 showing the "ACTIVE
SHUT" (top) and "ACTIVE OPEN" (bottom) states. The illustrations
are simplified by eliminating most of the components above bellow
508 level.
C. Pulsed Valves:
[0095] Certain advantageous implementations of valves in the art of
ALD manifolds and similar arts may be best accommodated by a
"pulsed valve" design. "Pulsed valve" is defined as a valve that is
used to introduce fluid from a delivery line into a chamber
preferably avoiding or minimizing a disadvantageous conduit between
the valve seat and the chamber. To better understand the definition
of "pulsed valve" and the distinction between pulsed valve and
conventional valve those who are skilled in the art are referred
FIG. 6 depicting a valve 500 with a conventional flow-path and, in
comparison, to FIG. 8 wherein a pulsed valve flow-path is depicted.
By comparison valve 500 depicted in FIG. 6 includes ports 540 and
550 with their respectively associated volume 545 and 555. The dead
space associated with conduits 545 and 555 is inevitable. In
contrast valve 600 depicted in FIG. 8 has only a minimized dead
volume at outlet 602 which is advantageous in certain applications
of gas delivery such as the pulsing of ALD effluents into an ALD
gas distribution module, such as an ALD showerhead.
[0096] The elimination of conduit 545 (FIG. 6) can also increase
attainable flow by increasing the over-all conductance between a
gas source and a process chamber. FIG. 8 depicts a pulsed valve
embodiment, 600, in accordance with the current invention. Supply
gas inlet 604 is linked to outlet 602 through valve chamber 607.
Hydroformed (or electroformed) bellow 608 seals over seat 610 when
the valve is either in the "INACTIVE" or the "ACTIVE SHUT" states.
Fail-safe mechanism includes stem 618, pad 620, actuator 624 and
other components that are not shown but can be deduced by those who
are skilled in the art by drawing similarly to the embodiments
presented in reference to valve 100 (FIG. 1a) and valve 200 FIG.
1c) above.
[0097] The major improvement exemplified in FIG. 8 is that the
volume of outlet 602 can be significantly minimized and the
conductance of outlet 602 is maximized. Additionally, valve 600 can
be easily integrated into the wall of a chamber or a showerhead to
achieve compact design and higher conductance. Conductance can be
increased, if necessary, by increasing the diameter of seat 610,
bellow 508 or both. In another close up look FIG. 9 depicts valve
600 shown in the "ACTIVE OPEN" (top) and "ACTIVE SHUT" (bottom)
states. The illustration is simplified by eliminating many
components above the level of diaphragm 508. However, the
embodiment exemplified in FIG. 9 corresponds to the FSNC design
depicted in FIG. 8.
D. Innovative Seals:
[0098] In the art of ultrahigh-purity valves a variety of valve
seat materials and shapes are known and successfully implemented.
Most commonly used materials vary in their properties such as
elasticity, tensile strength, impact resistance, hardness, chemical
compatibility, porosity, integrity and compression-set.
[0099] Typically, polymer based seals were preferred over elastomer
based seals due to their reduced porosity and minimized
deformation. However, some seal materials, in particular polyimide
based polymers and Ryton.TM. PPS polymers that are commonly
implemented for high-temperature applications are significantly
hard requiring increased sealing force to maintain the valves in
leak-tight "SHUT" states. The adverse consequence of using harder
valve seats and implementing tougher springs is significantly
shorter valve cycle lifetimes.
[0100] Commonly, valve seats in ultrahigh-purity valves were
implemented with shapes that incorporate significantly small radius
of seals at the portion facing the diaphragm. This design feature
was necessary to maintain leak-tight performance by increasing the
corresponding pressure at the contact of the seat with the
diaphragm. While these small radii have yielded better static
leak-tight performance, the smaller radii yielded accelerated wear
of both diaphragm and seat and reduced cycle lifetime. In
particular, serious deformation of seal and diaphragm were visible
shortly after the valves were put into service and continue to
develop throughout the valve cycle lifetime that could be linked to
the eventual failure. Most concerning, this mode of deterioration
was accelerated when the valves were actuated at high rates. While
polymeric seals were praised for their advantageous adaptation into
ultrahigh purity valves, they also constituted to the main cause of
failure and performance deterioration.
[0101] In contrast, perfluoroelastomers such as Kalrez, Chemraz E38
and the like can be utilized at high temperatures in excess of
250.degree. C. Elastomers are elastic but relatively soft and
require significantly smaller force to makeup a leak-tight seal.
Elastomers are notorious for a disadvantageous compression-set that
progressively accounts for disadvantageous deformation
under-pressure. Compression-set is accelerated at high temperature
and represents a significant difficulty for implementation as valve
seat seals. However, most perfluoroelastomers such as Kalrez 4079
and Chemraz E38 have shown a remarkable resistance to compression
setting beyond an initial .about.40% settings. Accordingly, these
elastomers are well-suitable for valve seat applications following
a fast "aging" process to enhance the inevitable 40%
compression-set under high pressure and temperature. Additionally
compression setting can also be bound by correct and restricting
design of the seal groove to restrict the deformation.
[0102] With a significantly reduced diaphragm-seat impact, FPV were
successfully implemented with relatively soft seals made from PFA,
PTFE and their equivalent. These perfluoro-polymers were
implemented within FPVs and withstood over 100,000,000 cycles at
220.degree. C. actuated at 10 cycles/sec with no deterioration of
valve performance. In contrast, PFA and PTFE seals are unpopular
within conventional high-purity and ultrahigh purity valves due to
their substantially high wear and deformation rate.
[0103] With the application of elastomers within valve seats, round
shaped seals can be implemented while spring-loading can be
significantly reduced. As a result, ultrahigh purity valve seals,
according to embodiments of this invention, have been useful to
suppress diaphragm and seal wear in fast actuating ALD valves that
were successfully implemented with valve cycle time as short as 10
msec.
[0104] Accordingly, elastomer based valve seats can significantly
extend the cycle lifetime of valves. However, elastomers contain
some residual porosity that may sorb and consequentially may leach
chemicals or other sources of contamination. In addition,
elastomers have limited resistance to abrasion.
[0105] In a preferred embodiment of the present invention,
high-purity FSNC valve 200 is implemented with elastomer seal to
improve valve cycle lifetime. Round cross-section seals are
implemented instead of the common small radii (tipped) seals. These
elastomer-based valve-seats are shaped appropriately to engage with
the valve seat. For example, FIG. 10a depicts such elastomer seal
622, that is pressure-mounted into the valve seat 623. The seal 622
forms a seal around the perimeter of the valve opening, and thus is
referred to as a perimeter seal. In the preferred embodiment it is
an O-ring shaped elastomer 622. Sometimes this may be referred to a
radial seal, as its location is at a certain radius about the valve
opening. The seal includes a ledge 628 extending perpendicular to
the plane of o-ring 622. The ledge matches a groove 626 in seat
624. The firm mounting of seal 622 into matching groove 626 is
useful in suppressing compression-setting deformation beyond a
pre-set 40% of accelerated deformation.
[0106] To improve the abrasion resistance of elastomer seals and to
eliminate porosity, elastomer seals are preferably coated with a
thin layer of polyimide or other polymer coatings that are known in
the art to be compatible with high temperature and harsh chemical
ambient. These films can be applied by a variety of methods
including dipping, spraying, spin-on, brushing, etc. A thin layer
of only several .mu.m is necessary. For example, Polyimide coatings
are available from HD Microsystems (Wilmington, Del.) in a variety
of properties and viscosities and are mostly suitable for
encapsulating perfluoroelastomer seals in according with
embodiments presented here. Specifically, PI-2545 which is solvated
in NMP (N-methyl-2-pyrrolidone) based solvent is most suitable to
enhance wear-resistance and inertness of Kalrez 4079 by
implementing a 2-10 .mu.m thick coating in 1-3 consecutive
application-cure cycles according to curing procedure provided by
the manufacturer. The resulting seal is highly suitable for valve
seal applications at 200.degree. C. Also recommended are
Bismaleimide coatings that are commercially available as BMI
adhesives from Polymeric GmbH (Berlin, Germany) and can be suitably
applied to encapsulate perfluoroelastomer seals after re-flow at
70-120.degree. C. At a temperature of .about.100.degree. C. these
BMI adhesives can be conveniently applied to form a thin overcoat
layer by dipping. Following a standard curing procedure
perfluoroelastomers coated with PX-300 type adhesives are suitable
for seal applications in the temperature range up to
.about.180.degree. C.
[0107] Other suitable coatings include FM2555 manufactured by Cytec
Engineered Materials, Inc. (Anaheim, Calif.) and materials such as
Durimide 7000 and Photoneece PWDC-1000 which are photoimagible but
can also be applied in the particular application described herein.
The polymer layer (634 in FIG. 10b) can be coated by a variety of
techniques including dipping, spraying, brushing and spin-on. The
polymer coated seals preferably preserve most of their elastic
properties. Some tuning up of seal resilience can be achieved by
the added layer of polyimide. Polyimide coated elastomer seals can
be mounted into the valve seat by using a ledge and groove such as
depicted in FIG. 10a. Additionally, adhesives such as PM2555 (Cytec
Engineered Materials, Inc.) or BMI PX-307 (Polymerics GmbH) can be
used to directly attach seal 620' onto an appropriately matched
rounded groove 632 in valve seat 624', as depicted in FIG. 10b.
[0108] In an additional improvement of seal performance and
resiliency this invention provides a metallic coating on top of the
polyimide (or other polymer) coating. Such coating is typically
applied using electroless plating to a thickness in the range from
0.0001-0.0020 inches. For the purpose of plating over the seal, the
seal is activated according to conventional plating over insulator
techniques that are known in the art. Following the activation, the
seal is coated with a pinhole-free metallic film such as Nickel or
Nickel-alloy. For example Nickel can be electroless deposited from
the sulfamate electrolyte as known in the art. Nickel alloys such
as Nickel/Cobalt, Nickel/Manganese, Nickel/Cobalt/Manganese and
Nickel/Iron, as well as combination of these alloys in a
multi-layer stack are useful to achieve good adhesion, low stress
and optimized mechanical properties. The elasticity and other
properties of the coated film are adjusted to provide an
hermetically sealing coating with appropriate resilience and
elasticity.
[0109] Additionally, methods that are commonly used to create
Electroforms over rubber mandrels are used to create an
encapsulating seal with sufficient strength that can encapsulate
the elastomer while good adhesion is not mandatory. In that case
the metallic electroform should be plated to a thickness typically
exceeding 0.001''. The coated seal is then mounted into the valve
seal. In an alternative approach the polyimide (or other equivalent
polymer coating, as disclosed above) coated seal is activated, then
mounted into the valve seat without metallic coating or with only a
thin layer of metallic coating and consequently the entire internal
area of the valve, including the seal is coated with a thin layer
of metallic film. This embodiment is illustrated in FIG. 11.
Elastomer (or polymer) seal 650 is coated with polyimide or the
like layer 652. Following, the seal is activated and possibly also
coated with a thin layer of metal 654. The seal is mounted into the
valve seat 656, as described above. Some of the area on the valve
is masked according to standard techniques to prevent metallic
coating, if necessary. Finally, the internal surface of the valve
is electroplated with a layer of metallic film, 658, such as Nickel
or Nickel alloy. Some of the plated film is also able to penetrate
into crevices 660 under the seal to fill-up these crevices and
provide additional attachment between the seal and the seat.
[0110] In another preferred embodiment, metallic seals are
electroformed and attached to the valve seat. Such seals, with
thickness in the range from 0.0010-0.0100 inches are made by
electroplating over an appropriate preformed body (mandrel) as
known in the art of Electroforming. For example, a useful
Nickel/Cobalt alloy called NiColoy.TM. available from NiCoForm,
Inc. (Rochester, N.Y.) can be applied to create such shapes by
electroplating over aluminum mandrels. FIG. 12 depicts an example
for metal seal Electroforming process and the subsequent mounting
of the preformed metallic-seal into a valve seat.
[0111] Accordingly, a seal shaped mandrel 700 is made with
dimensions appropriately scaled down to ensure the proper
dimensions of the final seal. For example if the final thickness of
the metallic seal is set to 0.0020 inches, the mandrel is made with
dimensions that are 0.0020 inches smaller in all directions. The
mandrel is preferably made from a metal that is easily plated and
easily dissolved away, such as aluminum which can be dissolved in
caustic solution as known in the art of electroforming. The mandrel
is shaped with a ledge 701. In subsequent step the mandrel is
clamped to prevent plating over the ledge using a two piece mask
702 and 704. The mask can be either conductive to facilitate
electroplating or nonconductive if electroless plating is to be
used. Subsequently, a metallic film 706 of appropriate properties
is plated over the mandrel. Seal properties can be tailored by
selecting appropriate metal alloy or metal alloys combination for
either a single or a multiple layer electroform. Following the
removal of the mask the ledge is exposed and the mandrel is
entirely etched away to create the seal 710 which is substantially
replicating the shape of the mandrel and has also a radial opening
wherein the ledge was located on the original mandrel. At this
point the resiliency of seal 710 can be improved, if necessary, by
a thermal hardening treatment, as known in the art.
[0112] In the next step, seal 710 is placed into an appropriately
shaped groove 714 in valve seat 712 and preferably brazed, welded
or glued in place. The brazing, welding or gluing must seal the
access into the opening of the seal. Preferably, a brazing process
is carried under elevated pressure of inert gas to determine the
resiliency of the seal by determining a preset pressure of trapped
inert gas inside the seal. Following the brazing step the seal is
preferably plated together with the internal surface of the valve
to create film 716 with advantageous filling into crevices 718
between the seal and the seat as described above. The resulting
seal uniquely integrates with the valve seat to provide the highest
standards of ultrahigh purity by eliminating some residual porosity
of polymers as well as the crevices between seal and seat. The
metallic seals can be designed by virtue of material and plating
process selection and thickness and shape selection to have the
appropriate combination of elasticity and robustness. Metallic
seals are mostly advantageous for high temperature
applications.
[0113] In another example FIG. 13 depicts a process flow to create
and mount a metallic seal that is advantageously radially open at a
specific angle to the plane of the seal. For example, an opening
that is centered around a 45.degree. conical plane as depicted in
FIG. 13. Mandrel 750 is shaped with a radial ledge 752 and a disc
754 wherein a hole 756 is formed. An insulating mask is made from 2
matching pieces 760 and 762 to match mandrel 750 and a contact
electrode 758 that is placed in good electrical contact with
mandrel 750. Preferably mask pieces 760 and 762 are made from a
relatively soft complying material such as rubber or Teflon to
provide hermetic seal over the masked parts 752 and 754 of mandrel
750. Following the masking, the mandrel is electroplated to created
film 764 over the exposed area of the mandrel. Then the mask is
removed and the mandrel is etched out to create the seal 764. High
volume manufacturing of seals 764 is accomplished by mounting
multiple numbers of mandrels 750 on a single mask 760+762 shaped
appropriately to accommodate multiple mandrels.
[0114] Seal 764 is placed in contact with valve seat 766 over an
appropriately shaped outward pointing radial corner 768.
Subsequently the seal is brazed or welded to the body of the valve
as schematically described by 772 and 772'. During the brazing or
weld process, an appropriately pressurized inert gas 770 is
entrapped inside the seal. If welding is desired, the shape of seal
764 must include some end ledges to conform to the requirements for
welding. For example, the specifications for electron-beam welding
from Servometer Precision Manufacturing Group, LLC (Cedar Grove,
N.J.). In the case shown only the edge 772 can be done by welding,
preferably electron beam welding. Following, the valve is plated to
created top coating 774 as described before. In a preferred
implementation the material and plating process to create 764 is
selected to provide best properties of elasticity, weldability etc.
while the properties of film 774 can be independently selected to
achieve best chemical resistance and sealing properties. For
example seal 764 is preferably made from a double layer of High
Hardness NiColoy.TM. followed by a layer of Ni/Co/Mn alloy.
Accordingly, high tensile strength can be reproducibly achieved, on
the order of 140 GPa, without the need for thermal hardening.
Electroformed NiColoy.TM. exceptionally maintains over 95% of its
tensile strength at 300.degree. C. In contrast, a softer film such
as Nickel is preferred for film 774 to create a less brittle and
more complying surface.
[0115] The metallic seal and, in particular, the integrated
metallic seal taught in this invention is very useful in creating
low profile valve seats. For example, FIG. 14 compares the
seal-seat arrangement, 800, that was depicted in FIG. 13 with
arrangements 820 and 840 that are most suitable for pulsed valve
applications and advantageously have low profile. Seat 820 utilizes
a metallic electroformed seal 822 that matches with a corner-shaped
valve seat 824. Seal 822 is flexed into the seat and is placed to
fit over the corner 825. Brazing or welding 826 is used to attach
seal 822 as described above. During the attachment some
well-defined entrapped inert gas 830 at defined pressure can be
encapsulated into the seal. Following the attachment, the entire
seal-seat assembly is preferably plated with film 828. To improve
sealing properties, film 828 is preferably made of relatively soft
material such as Nickel. Alternatively, more than one layer is
plated within the process of laying down 828 such that both
resiliency and compliance is achieved. For example, a layer of
high-strength NiColoy.TM. is first plated for the thickness of
0.001 inches followed by a layer of Nickel to a thickness of 0.0005
inches. Additional embodiment 840 presented in FIG. 14 is also
applicable mainly for pulsed-valve applications.
E. Ultra High Purity Spring-loaded Bellow Valves:
[0116] Some applications, in particular the introduction of
low-volatility chemical vapor into an ALD reactor (for example
through a showerhead module) require valve with very large
conductance. Accordingly, the present invention teaches an
embodiment for FSNC valve wherein a spring-loaded bellow,
preferably a welded or a thin, electroformed below is implemented
within the valve chamber to provide the seal. For example, FIG. 15
depicts a schematic cross-sectional side-view of such embodiment.
Valve 900 is a high conductance pulsed valve capable of directly
injecting vaporized gas into a process chamber or a showerhead
through outlet port 918. A valve seat 916 is integrated into the
external wall 920 of the showerhead or the chamber. The seal of
seat 916 that is shown is similar to seal 820 (FIG. 14) and in the
shown case is integrated into the valve seat to create a very low
profile path 918.
[0117] Sealing plate 904, made of suitable metal, is welded at the
perimeter to bellow 902. The other end of bellow 902 is welded to
disc 906 that is made of suitable metal. The perimeter of disk 906
is mounted by the pressure of bonnet 930 and fastening nut 932 to
create the perimeter seal 908. Perimeter seal 908 creates the
sealed control chamber 931 and valve chamber 903. Valve chamber 903
is connected to chemical port 928 through conduits 926 and 922.
Control chamber 931 is used for valve actuation between the "ACTIVE
SHUT" (shown) and "ACTIVE OPEN" states. The backside of sealing
plate 904 is fastened to spring loading post 910. Spring 912 is
compressed when the valve is at "ACTIVE SHUT" state. When the
pressure is relieved from control chamber 931, spring 912
decompresses to move sealing plate 904 away from seal 916. The
spring is compressed between the end of post 910 and spring mount
905. Several holes 907 are formed in the wall of spring mount 905
to facilitate fluidic communication within the entire volume of
control chamber 931.
[0118] The FSNC mechanism includes stem 934 and pad 936. Bellow 938
is used to allow stem vertical motion while keeping the control
chamber 931 leak tight. Bellow 938 is utilized for very high
temperature applications wherein sliding elastomer seals on the
stem might be inadequate. Alternatively, bellow 938 is replaced
with a sliding elastomer seal for lower temperature applications,
as commonly done in the art. In high temperature applications, stem
934 and stem guide 941 are preferably extended to provide
sufficient separation between the valve and the actuator. This
separation allows to maintain the valve at temperature
substantially higher than the actuator. Alternatively, as depicted
in FIG. 17, the actuator can be made compatible with higher
temperature, eliminating the need for spatial separation. Actuator
942 is equipped with conductance adjustment mechanism 944 that
allows for externally tuning the limit for stem 934 motion. When
the actuator is driven to the "ACTIVE" state by pressurizing
actuator 942 through port 950, the stem moves away from the valve
until it is stopped by limiting mechanism 944. The position of pad
936 is determined by the range of motion of stem 934. The range of
opening for sealing plate 904 is determined by the place of pad 936
wherein pole 910 is stopped when the valve is actuated into the
"ACTIVE OPEN" state. Accordingly, the conductance of valve 900 is
determined externally by adjusting the position of 944.
[0119] Valve 900 is supplied with compressed air or inert gas
through conduit 946. The air supply splits in tee 948 into actuator
feeding line 950 and pilot valve feeding line 949. Pilot valve 952
is a normally open solenoid valve and provides a path for air to
pressurize the control chamber for maintaining the valve at "ACTIVE
SHUT" state when the pilot valve is not actuated. The air is
supplied through feeding line 956 and port 958. When the valve is
actuated into the "ACTIVE OPEN state pilot valve 952 is energized
to seal the air path to 949 and to link control chamber 931 to the
vent/evacuation port 954. FIG. 16 illustrates valve 900
schematically in the "ACTIVE SHUT" (bottom) and the "ACTIVE OPEN"
(top) states. For the sake of simple illustration, most parts above
the level of bonnet 930 are not shown. The conductance of valve 900
in the "ACTIVE OPEN" state is determined by the opening gap 970.
Gap 970 is determined by the possible translation, of the bellow
assembly to make contact with pad 936 (not shown here, shown in
FIG. 15). Accordingly, the position of pad 936 determines the stop
position 980. As described in reference to FIG. 15 stop position
980 is externally adjustable through mechanism 944 (FIG. 15).
[0120] Also shown in FIG. 15 are cavities 924 and 922. In some
embodiments these cavities comprise of a round slot that was
machined into the wall 920 as an annular cavity wherein 922
represents the left side and 924 represents the right side of the
annular cavity, respectively. Following, the angled conduit 926 was
drilled to link into the edge of 922 prior to the welding or
brazing area of conduit 928. This arrangement provides a high
conductance path for low pressure gas into the valve chamber 903.
Gas continuously fills chamber 903 and the other linked spaces 924,
922 and 928 and is available for pulsed delivery through the
opening 918 when the valve is actuated into the "ACTIVE OPEN"
state. The translatable bellow assembly of 904, 910 and 902 has a
relatively small mass that enables down to sub-millisecond response
for valve cycling and low damage actuation. The convoluted surface
of the bellow does not pose contamination issue if the bellow is
prevented from approaching full compression. Accordingly, valve 900
maintains standards of ultrahigh purity. Additional improvement
adds a thin layer of polyimide coating or other suitable polymer
over the sealing plate 904 to prevent the generation of particles
from a metal to metal contact between plate 904 and seal 916. It is
also useful to coat bellow 902 with a suitable metallic film and/or
polymeric film to improve chemical inertness and cleanliness, when
necessary.
F. High Temperature Valves:
[0121] In recent years there is a growing need for reliable valves
that can operate at high temperatures, up to 200.degree. C. and
beyond. In particular, ALD of many useful materials has been
restricted to low-volatility chemicals wherein useful vaporization
of desired chemicals required high-temperature valve manifolds. The
challenges of high temperature valves are several folds: [0122] a.
High-temperature compatibility and properties of seal materials.
[0123] b. High temperature elasticity of diaphragm and spring
materials. [0124] c. Chemical compatibility and purity of wetted
surfaces. [0125] d. High temperature compatibility of pneumatic
actuators. [0126] e. High temperature compatibility of pilot
valves.
[0127] Embodiments in accordance with this invention that were
described above provide innovative seals with extended temperature
compatibility. For example, elastomer seals made from Kalrez,
Chemraz and the like can reach continuous operating temperature of
260.degree. C. Perfluoro-polymers such as PFA and PTFE were proven
to be suitable as FPV seals with exceptional performance and cycle
lifetime exceeding 100 million cycles at 220.degree. C. Many
polymeric coatings such as polyimide PI-2545 and PX-300 BMI
adhesives have high temperature compatibility and are useful to
coat the elastomer seals for improved seal performance as discussed
above. Most suitable for high temperature operation are metallic
seals that were disclosed above. These metallic seals can be used
for operation temperature in excess of 300.degree. C. provided that
the proper selection of materials prevents corrosion and possible
contamination.
[0128] The properties of diaphragm materials, in particular
elasticity and corrosion resistance are important selection
criteria for high temperature applications. For example
heat-treated-quenched Hastelloy C276 maintains .about.86% of its
elasticity at temperature exceeding 500.degree. C. and .about.92%
of its elasticity at .about.300.degree. C. Likewise, Inconel 603XL
maintains .about.93% of its elasticity at .about.300.degree. C. and
.about.90% of its elasticity at .about.400.degree. C. Similarly,
Inconel 706 maintains .about.88% of its elasticity at 300.degree.
C. Other alloys that are compatible with high temperature
applications include Nimonic alloys such as Nimonic 90 and Heat
treated Titanium alloys such as ST40. Elasticity based selection
criteria apply also for materials suitable for high temperature
implementation of rippled diaphragm materials and electroformed and
hydroformed bellows materials in accordance with embodiments of
this invention.
[0129] Bellow valves in accordance with this invention are mostly
suitable for high temperature applications provided that corrosion
resisting materials and welding procedures are maintained. The
implementation of return spring 912 (FIG. 15) should follow the
correct selection of high temperature maintained elasticity as
described above in reference to diaphragm material selection.
However, since spring 912 is not in contact with the delivered
chemicals, contamination, corrosion and oxidation issues are not
important and the selection of spring materials is broader.
[0130] Selection of proper materials for high temperature
diaphragm, electroformed bellows, hydroformed bellows and welded
bellows should also follow chemical compatibility at those high
temperatures. With this respect Inconel type alloys, Titanium
alloys and Hastelloy offer a broad coverage of chemicals with
corrosion and oxidation resistance maintained at temperature
exceeding 300.degree. C. In certain cases, improved chemical
compatibility at elevated temperature is achieved by plating the
diaphragm or bellow with metal such as Nickel or coating with a
thin-adherent polymer film.
[0131] High-temperature compatibility of pneumatic actuators must
be considered, as well. For example, standard pneumatic actuator
utilized seals and grease which cannot exceed operating temperature
of .about.80.degree. C. However, standard actuators can be easily
upgraded with high-temperature elastomer seals and high temperature
lubrication to be operational up to .about.250.degree. C. Beyond
250.degree. C. valve operation temperature, the actuators can be
placed remotely from the valve to maintain a temperature gradient
and cooler actuator as described in reference to FIG. 15. In this
implementation, the seal over control chamber 931 is maintained
with a metallic bellow 938. Alternatively, actuators can be
upgraded for operational temperature beyond 250.degree. C. by
replacing elastomer seals with bellows as depicted in FIG. 17.
Accordingly, the actuator can be made enclosed with bellow 1014 and
be actuated by inflating bellow chamber 1020 to create a pull force
over stem 1002. High temperature compatible alloy materials must be
selected for spring 1008 but corrosion and contamination are not an
issue since spring 1008 is not in contact with the chemicals.
[0132] In high-temperature applications, the air or inert gas
supplied to the valve is preferably preheated to avoid localized
cooling of the diaphragm or bellow by air introduction into control
chamber 1009. This is an important part of the method taught in
this invention and those who are skilled in the art can implement
suitable reservoir of heated air to ensure that pneumatic cooling
is not a performance limitation. Additionally, care should be taken
to ensure that the pilot valve, typically an electromechanical
solenoid valve, is compatible with the high temperature.
G. ALD Manifold with Integrated Pulsed Valves:
[0133] ALD manifolds are preferably implemented with pulsed valves.
This implementation supplies the inert and reactive gasses that are
needed for ALD with minimized delay and cross contamination and
enables highest conductance valve-implementation, when necessary.
FIG. 18 illustrates schematically an ALD injection system 1100
comprising of 4 valves. Valves 1140 are integrated into the top
wall of a showerhead with internal space 1120 and a nozzle array
1130. Baffle disks 1125 are typically mounted within space 1120
across from the opening of valves 1140 to avoid localization of
flow insertion through the "line-of-sight" nozzles across from a
valve 1140 opening. In the example of FIG. 18 valves 1140 are
bellows valves, similar to the embodiment described in reference to
FIG. 15. However, many different implementations are suitable and
can be appropriately selected to fit the temperature range and
conductance that are necessary. Also, within manifold 1100, valves
with different design for different ranges of conductance may be
implemented to optimize the manifold to the specific ALD process.
Typically, though, all the valves are maintained at the same
temperature, as well as the showerhead.
[0134] Under proper execution of ALD, gas-phase mixing of chemicals
in showerhead space 1120 is substantially avoided. However, some
film growth of up to a monolayer per cycle occurs on exposed area
of sealing plate 1142 and seal 1144. This growth can be minimized
by appropriate selection of showerhead temperature. Breaking a
bridging film between sealing plate 1142 and seal 1144 could be a
source of particles generation. However, the growing film cannot
bridge between the seals and the sealing plates (or, for example,
the diaphragm in other implementations) since it is disconnected
every cycle by virtue of valve actuation. Yet, the prospect of film
peeling from seal 1144 due to the flexing of that seal requires
considerations and measures to avoid source of particulates
contamination. These measures include minimizing the elasticity of
the seal by, for example, reducing the diameter of the seal.
Additionally, the exposed surface of the seal is slightly roughened
by mechanical abrasion, etching or both to improve adhesion of
deposits and reduce the stress in these growing deposits.
H. Optimization of Diaphragm Based, Fast FSNC Valves
[0135] Adaptation of ultrahigh-purity confined valve designs into
embodiments disclosed in this invention requires that measures are
taken to ensure appropriate functionality of metallic diaphragms
under fluid control. Specifically, conventional dome-shaped
diaphragms must be mounted or otherwise reshaped to ensure
leak-tight seal under "ACTIVE SHUT" state as described with
reference to embodiments of this invention. FIG. 19 illustrates the
prior art mounting of a domed shaped diaphragm (19a) that is
substantially inadequate for practicing the invention, the generic
solution (19b), the complementary seat redesign that is also taught
by this invention (19c) and an advantageously implemented rippled
diaphragm (19d).
[0136] Dome-shaped diaphragms are readily available and easily made
out of many different materials, as described above. For many
applications that do not require large conductance. beyond
C.sub.v.about.0.5 dome-shaped diaphragms are highly recommended and
well suitable. However, these diaphragms tend to disadvantageously
distribute sealing force into a large area when actuated by
pressurized fluids. As a result, sealing over conventionally
designed valve seats is mostly inadequate. A slightly modified
diaphragm, wherein a reshaped perimeter is implemented, is taught,
and is depicted in FIG. 19b. The modified diaphragm includes a
somewhat flattened dome 180 (largely exaggerated in FIG. 19b for
better clarity) that provides extra-stiffness at the perimeter of
the diaphragm. In another useful embodiment, standard dome-shaped
diaphragms are deformed by mounting the diaphragm under fluid
pressure applied from the high-purity side of the diaphragm
(description is given below in reference to FIGS. 20a and 20b).
Additional improvement of valve performance is gained by increasing
the diameter of the seat as depicted in FIG. 19c. The increased
diameter seat is also useful to increase the conductance of valve
100. Additional reinforcement of diaphragm perimeter by adding a
radial ripple, 184, as shown in FIG. 19d further improves fluid
control while maintaining compatibility with stem actuation.
[0137] Diaphragm mounting with suitable pre-set elastic-deformation
(as a result of pre-set radial-deformation) is useful to create a
preloaded focusing force within the diaphragm. This pre-set stress
can be implemented by a suitable diaphragm mounting-method.
Accordingly, a diaphragm fastening method wherein the diaphragm is
lightly-tightened into place, for example by applying a torque of
10 N.times.m (Neuton.times.meter), followed by fluid pressurization
into the valve chamber (from the high-purity side of the diaphragm,
110') is utilized to flex the diaphragm backwards prior to the
final fastening, for example with a torque of 70 N.times.m, and
obtain the preferred deformation depicted schematically in FIG.
19b. When the diaphragm 108 is pressurized from the valve chamber
side (pressurizing valve seat inlet 110') while the edge 109 is
still not completely secured, edge 109 slides inwards and the
diaphragm is deformed into a slightly perimeter stiffened shape,
180, once the edge 109 of diaphragm 108 is finally secured by the
pressure of bonnet 112 and the pressurizing gas (185 FIG. 20b) is
removed. The deformed diaphragm has a radial pre-set higher stress
area 182, localized substantially at the perimeter. This procedure
toughens the diaphragm at the perimeter and, in turns, causes the
diaphragm to invert, under pressurized fluid force (from port 116),
from the center outwards. As commercially available, or otherwise
suitable, diaphragms vary in thickness and spring constant, the
mounting procedure should be optimized by a design of experiment
(DOE) procedure that, per seat and diaphragm design includes 3-4
different values of pressure used for diaphragm 108 mounting and
for fluid actuation. For example a pressure between 35 to 55 psig
was found useful for pre-setting a 0.0020'' thick, .about.1''
diameter Elgiloy diaphragm for optimized fluid actuation (that is
compatible with stem actuation, as well) in the range from 30-100
psig. It was also found empirically that pre-set diaphragms do not
typically pre-set permanently by the mounting procedure, described
herein.
[0138] Additionally, rippled diaphragms are very useful for
presetting advantageous localized stiffness into diaphragms. For
example, the rippled diaphragm depicted in FIG. 19d is suitably
optimized for both stem and fluid actuation, in accordance with
this invention. Accordingly, radial ripple, 184, adequately
reinforces the perimeter of the diaphragm 108. As discussed above
rippled diaphragms can be readily hydroformed or otherwise
electroformed into any useful shape as known in the art.
[0139] FIG. 20 provides a breakdown of a dome-shaped diaphragm
mounting procedure. In FIG. 20a diaphragm 108 is adequately placed
between valve 102 and bonnet 112, stem 118 is actuated and pilot
valve 144 (not shown) is actuated to the "ACTIVE OPEN" state.
Following, the diaphragm is lightly secured, in place, using nut
111 and a torque wrench set with a small torque chosen in the range
from 5-15 N.times.m. Then in FIG. 20b valve port 106 is sealed
using plug 190 and gasket 188. The port 104 is then pressurized by
clean gas 185 to a pressure of 45 psig. The gas source must
maintain the desired pressure (45 psig, in this example) even
though there is a slight leak at the diaphragm perimeter 109, since
diaphragm 108 is not completely secured. Diaphragm 108 settles into
the pre-set shape practically instantaneously, following the
introduction of gas 185. At this point the diaphragm edge is
slightly pulled inwards by the fluid applied deformation and the
center of the diaphragm is slightly dimpled by the button 120.
Finally, nut 111 is used to completely and adequately secure
diaphragm 108 using, for example, a torque of 70 N.times.m. Once
the pressure 185 is removed, diaphragm 108 flexes back into a
"free-standing" shape. Since the perimeter of the diaphragm was
pulled inwards by the mounting procedure (compared to the
stand-alone diaphragm), the "free-standing" shape of the diaphragm
is slightly radially deformed.
[0140] While it is adequate to adapt standard ultrahigh purity
valve seat and diaphragm designs to perform well in embodiments
taught in this invention, using the back-pressure mounting
procedure, as described herein (FIG. 19b), it is further
recommended in the preferred embodiment, herein, to further improve
both design versatility and valve reliability by a 30-100%
increased seal 110'' diameter, as depicted schematically in FIG.
19c (as compared to FIG. 19b).
[0141] The descriptions and examples of the preferred embodiment
further explain the principles of the invention and are not meant
to limit the scope of invention to any specific method or
apparatus. All suitable modifications, implementations and
equivalents are included in the scope of the invention as defined
by the following claims.
* * * * *